An increasingly number of jurisdictions are making use of civil recovery (termed civil forfeiture in the United States) as a weapon to seize property that is linked to crime. A key feature is that the holder of the property need not have been convicted of having committed the crime to which the property is said to be linked – or indeed of any crime at all. This gives rise to the question: what protection, if any, should be given to the holder where there is no evidence that they knew or even suspected that the property was linked to crime? Different jurisdictions have different answers to this. It is therefore argued that a balance needs to be drawn. The mechanism for seizing the proceeds of crime – and possibly also the instrumentality of it – must be effective, but equally, this should not mean undue hardship for those who had no knowledge of the crime in question, let alone any part in it. Finding an appropriate balance may not be easy, but it is important. The attempt by the United Kingdom to strike such a balance has not been satisfactory, as the English High Court confirmed in National Crime Agency v Azam (No. 2) (2014), and therefore does not provide a useful model. The paper recommends an alternative balance which may be recommended to jurisdictions as they continue to develop their own provisions.