Hard to Handle: A Look at Hard vs. Soft Data

The "pandemicycle" (yes, that's a new term from us) has been unique in myriad ways; including multiple divergences within the economy. Divergences have occurred within the economy between goods and services, cyclical and noncyclical segments, and discretionary and nondiscretionary categories of spending; among others. A divergence we're tackling today is between soft and hard economic data. First, the definitions:

  • Soft economic data refers to surveys, sentiment indicators, and expectations, such as consumer confidence, business outlook surveys, and Purchasing Managers' Indexes (PMIs).
  • Hard economic data refers to measurable and objective metrics like gross domestic product (GDP), employment readings, retail sales, and industrial production.

A widely watched metric from Bloomberg is its Economic Surprise Index which tracks the difference between actual data releases and economists' forecasts. Bloomberg also separates them into hard data and soft data components, shown below.

Surprise!

Bloomberg hard and soft data surprise indexes

There are a couple of interesting periods over the past decade, as shown above, including the surge in soft data surprises in 2017 (typically seen as tied to President Trump's win in late 2016). While hard data surprises did start to lift in late 2017, they remained well below soft data surprises throughout 2017 and 2018. Fast forward to the pandemic, another stark divergence opened up—again in favor of soft data surprises—as the economy started to find its footing in the latter part of 2020 into 2021.