Tackling Margin Pressures on the Mortgage Book: A Data-Driven Solution

June 19, 2019

Mortgage Pricing

It’s hard to ignore the irony behind TESCO’s recent exit from the mortgage market – driven out by a price war that led to compressed margins. Margins have been squeezed for lenders across the board due to increased competition. The likes of HSBC have expanded their mortgage book by about 10% year on year. The prevailing interest rate environment and looming Brexit uncertainty meant TESCO could not afford to earn sub-par margins for much longer.

Given these tough conditions and the commoditisation of the mortgage market, banks can improve margins by taking a more data-driven approach to pricing their mortgage books.


What do banks do wrong in pricing mortgages?

Too often product and pricing teams within banks rely on intuition and individual judgement to make pricing decisions. Consequently:

  • Valuation models have sub-optimal assumptions making mortgage pricing inflexible
  • Key factors like volume response to ‘rank’ based on historical trends are not considered
  • Pricing is not thought of at the book level leading to decisions that are not consistent with the overall goal of maximising the objective function (e.g.: NPV)


Lenders should take a more data-driven approach to mortgage pricing

A five step approach to price optimisation:

  1. Ask strategic questions and generate consensus amongst key stakeholders
    • What is the right framework to think about volume v. margin trade-offs?
    • Is there appetite to differentiate acquisition vs. retention pricing?
    • Can we break new ground by bundling certain features into mortgage products and charging them based on customer preferences
  2. Model price elasticity of your customers. Pricing for mortgages should consider both price as well as non-price factors from multiple sources to correctly understand customer behaviour. For instance: understanding vintage of customers can be useful in pricing certain products
  3. Identify multiple functional forms (e.g.: linear, cubic, logarithmic, etc.) to create regressions on selected variables. Select the right models for the right product sets based on appropriate statistical criteria
  4. Create product maps by clustering similar products in order to make your models flexible
  5. Create intuitive front ends that product managers can use to optimise pricing for the entire book in one go using optimisation algorithms


Key benefits to data-driven mortgage pricing

Using data-driven techniques in pricing can have a deep impact on the bank’s profitability and also lead to more robust processes and governance.

  • Data-driven models can consider a vast array of variables in a dynamic manner that enables models to adapt to changing conditions. Using statistical modelling techniques helps product and pricing managers make considered decisions by framing complex decisions like volume vs. margin trade-offs in the right context
  • Data driven approach also facilitate effective governance structures through audit trails which can be baked into the process if using suitable tools
  • Price optimisation techniques based on a deep understanding of customer elasticity can help improve margins for mortgage products

In a market experiencing margin pressures, lenders who use advanced data-driven techniques in pricing are likely to win what can be a long war of attrition.