Guanxi-based resource allocation is a prevalent phenomenon in China. Previous research on guanxi-based resource allocation has at least two limitations. First, guanxi exercisers and guanxi beneficiaries are two main actors in a guanxi-based resource allocation decision. However, past research mainly focused on effects of "guanxi" factors between guanxi exercisers and beneficiaries on resource allocation results, and most of them neglected effects of "non-guanxi" factors (e.g., performance of guanxi beneficiary). Second, few research examined when decision-makers adopted guanxi-based resource allocation decisions and when they neglected guanxi factors to allocate resources. Based on the above concerns, this study examined effect of guanxi beneficiary's performance on guanxi exerciser's resource allocation and explored the processes of guanxi-based resource allocation under Chinese cultural background. Based on existing theories, the present study proposed that guanxi beneficiary's performance was negatively related to over-allocated resources that guanxi exerciser gave to guanxi beneficiary, and that guanxi beneficiary's performance was negatively related to inequity of guanxi exerciser's resource allocation. In order to examine the above hypotheses, the present study designed four within-group situation-based experiments. In these experiments, participants were asked to play as a resource allocation decision-maker. They needed to allocate resource to guanxi beneficiary and another partner based on their performance. They were supposed to have strong guanxi (experiment 1, 3 and 4) or moderate guanxi (experiment 2) with the guanxi beneficiary, but have moderate guanxi (experiment 3) or weak guanxi (experiment 1, 2 and 4) with another partner. Guanxi beneficiary’s performance was manipulated at three levels, which were low performance (i.e. guanxi beneficiary’s performance was much lower than another partner’s performance), equal performance (i.e. guanxi beneficiary’s performance was equal to another partner’s performance) and high performance (i.e. guanxi beneficiary’s performance was much higher than another partner’s performance), respectively. Results of four experiments repeatedly verified our hypotheses. When guanxi beneficiary’s performance was significantly lower than another partner’s performance, guanxi exerciser would consider guanxi factor with guanxi beneficiary and allocate more inequitably extra resources to guanxi beneficiary, which significantly exceeded the resource that they deserved. When guanxi beneficiary’s performance was significantly lower than another partner’s performance, however, guanxi exerciser would not consider the guanxi factor with guanxi beneficiary and equitably allocate resource to guanxi beneficiary and another partner. These findings suggested that only when people were in an obviously disadvantageous situation, guanxi would play a compensatory role in resource allocation processes. Our findings have some implications for guanxi-based resource allocation theory and management practices. First, our findings explore the boundary conditions of guanxi-based resource allocation theory and suggest that guanxi does not always have effects on resource allocation in Chinese context. Second, our findings suggest that managers not only need to encourage employees to maintain good guanxi with them, but also need to encourage these employees to improve their performance, which could decrease the pressure to make guanxi-based decision. Moreover, our findings also suggest that managers could recruit employees with good guanxi and high performance, but they need to avoid recruiting employees with good guanxi but low performance.