The study attempts to examine herding behaviour in Pakistan stock exchange and determine how herding behaviour responds to asymmetric market conditions. This study has employed CH model, CSSD & CSAD models, and State‐space model. We have used time series data for the period from 2000 to 2016, which include daily returns of KSE all share index and 890 firms listed on Pakistan stock exchange. The outcomes of the study not only reported significant evidence of herding behaviour in Pakistan stock market over the entire sample period but also found it to be more pronounced under extreme market conditions, market volatility and financial crisis. Further, it was found that herding behaviour increases before the crisis, whereas decreases at the time of crisis. Fund managers and investors need to take herding behaviour in to account to evaluate asset prices and require a large number of stocks for achieving a correlation of inferior degree in portfolios and same degree in diversification. In addition, funds herding can be used as a signal of managerial quality. Moreover, as herding implies that fund managers and investors react to regulatory changes in a similar way, which may prevent asset prices from reaching their fundamental values and further causes financial markets to become fragile and destabilize, thus policymakers should take measures to control herding.