Automated underwriting is a system by which a computer makes a decision about whether or not a loan or other financial transaction should be underwritten. To underwrite a transaction is to guarantee that one party will get the expected money, a service provided in return for a commission. With automated underwriting, a set formula decides whether this service will be offered.
There are a vast range of financial transactions that can be underwritten. This could include a simple loan, with the underwriter taking on the risk that it will not be repaid. It could be a stock issue, with the underwriter taking on the risk that not all the shares will be bought by the public. One slight variation is with insurance underwriting, which is simply the process of issuing a policy. In this case, the insurance firm is effectively underwriting by taking on the risk that would otherwise be borne by the customer.
With loans, including mortgages, automated underwriting does not necessarily refer to underwriting in the traditional and strict sense of the word. It often refers simply to a computerized service that gives an "objective" decision of whether the loan should be made. The creators or operators of the service may provide it on an advisory basis only and will not necessarily guarantee that the loan will be repaid. As the lender retains the risk, the fee paid to use the computerized service will be considerably lower than if the service providers were literally underwriting the loan.
The main benefits of automated underwriting are speed and objectivity. Systems can process an application and issue a decision in seconds, without the need for a human to manually assess each piece of information. The system will also work to fixed rules and make a rational decision, as opposed to a human who might make a gut decision based on their personal interaction with the prospective borrower. Of course, this could be seen as a disadvantage as it makes the process more impersonal and removes judgment.
The main drawback of automated underwriting is inflexibility. For example, they may automatically reject an application from someone with a poor credit history even if a previous default can be explained and might be subjectively assessed as being unlikely to indicate future defaults. An automated underwriting system may also give a "black or white" decision where there is no distinction between a customer with an atrocious credit risk and a customer who very narrowly missed out on the automated acceptance, where a human underwriter might have been prompted to examine the circumstances and make a judgment call.