Hardy presented a famous argument that shows that there exist no hidden variables in quantum mechanics. His argument is called Hardy's paradox. The commentaries and discussions on Hardy's paradox have been ordinarily presented for deterministic hidden variables, and it has not been discussed whether a similar paradox will occurs in probabilistic hidden variables. So, first, I will show that Hardy's argument can also be applied to probabilistic hidden variables. After that, I discuss the common cause of a correlation, which has been actively discussed in the philosophy of science in recent years. That is because requiring the common cause of a quantum correlation is mathematically equivalent to introducing a probabilistic hidden variable to explain the correlation. Some researchers argued that in the non-existence proof of hidden variables, the existence of a common common-cause to multiple correlations is tacitly assumed, however that is unreasonably strong. According to their idea, it is sufficient if there exists a respective common-cause of each correlation. However, it will be revealed that at least as far as Hardy's paradox is concerned, merely weakening the common common-cause requirement to the respective common-cause requirement does not help to construct a local model.