Cross-sectoral dispersion in firms’ earnings expectations during the COVID-19 crisis

Prepared by Joost Bats, William Greif and Daniel Kapp

Published as part of the ECB Economic Bulletin, Issue 5/2021.

While the recovery in equity prices over the last year has been underpinned by improvements in aggregate short-term earnings expectations (Chart A), developments in the latter have varied greatly across sectors (Chart B). Compared with the situation before the pandemic, expected earnings remain persistently subdued in sectors such as travel and tourism, while growing rapidly in the technology, utilities and non-bank financial services sectors. As a result, current earnings expectations are in line with what is called a cross-sectoral K-shaped recovery (where the paths of the best and worst-performing sectors resemble the two arms of the letter 'K'). Similar dispersion holds true across countries (see also Box 3 of this issue of the Economic Bulletin and the May 2021 Financial Stability Review), in line with the idea that some countries are economically more dependent on particular sectors than other countries.

This box measures the heterogeneity of analysts' daily 12-month ahead earnings per share (EPS) forecasts for listed firms in 20 different sectors. As there is always a possibility for analysts to revise their estimates when new information becomes available, earnings forecasts are especially useful for gauging the immediate effects of the COVID-19 crisis on cross-sectoral dispersion.

To capture the impact of the COVID-19 crisis on cross-sectoral dispersion in earnings expectations, the Gini coefficient is used. For a variable of interest (here: EPS forecasts) sampled across individuals (here: sectors), the Gini coefficient provides a scaled average of the absolute differences in the variable's outcome across all pairs of individuals.

As such, it is an indicator of relative dispersion ranging from 0 to 1, where an increase in the Gini coefficient reflects an increase in dispersion.

The Gini coefficient has several advantages over simpler measures of dispersion such as standard deviation. First, it represents the degree of inequality in any statistical distribution. Second, the Gini coefficient judges the dispersion through a relatively easy-to-interpret summary metric: it reaches its maximum value of 1 when earnings expectations are positive for one sector and zero for all others, and it is 0 when all sectors contribute equally to the sum of expected earnings. In addition, in this box we calculate the Gini coefficient using earnings expectations data based on levels instead of growth rates, so as to assess whether changes in cross-sectoral dispersion are structural.

The data show that, according to this measure, cross-sectoral dispersion in earnings expectations has persistently risen since the onset of the COVID-19 crisis, in contrast to past crises where the impact was more homogeneous across sectors (Chart C). While the Gini coefficient dropped after the collapse of Lehman Brothers during the Global Financial Crisis - reflecting a broad-based downward revision of the earnings outlook across sectors - it has jumped upwards since the start of the COVID-19 crisis. Together, these observations signal a structural shift in expectations during the current pandemic: the market expects some sectors to persistently underperform others over the coming year.

The increase in social distancing and the associated decline in mobility due to lockdown measures help to explain why the dispersion in earnings expectations has risen during the COVID-19 crisis. The reason is that mobility restrictions affect economic activity differently across sectors. The correlation between the stringency of lockdowns and the cross-sectoral dispersion in earnings expectations is estimated by means of local projections using daily data from early January 2020 to late April 2021.

As an indicator of lockdown stringency, the Goldman Sachs Effective Lockdown Index (ELI) is used. The ELI accounts for a combination of official government restrictions and actual mobility data.

This empirical analysis shows that cross-sectoral dispersion in 12-month EPS forecasts has risen each time lockdown measures have further restricted mobility (Chart D). Up until December 2020, imposing a lockdown such that the ELI would rise by 50 points (which is half of its maximum range of 0 to 100 and corresponds to the change observed in March 2020) increased the Gini coefficient by 3 percentage points after 20 working days.

At the same time, the implementation of vaccination campaigns has been a game changer: stringent lockdowns have added far less to the dispersion metric since vaccination started in late 2020 (Chart D).

Since mid-December 2020, when vaccination started in the euro area, the effects of restrictive lockdown measures on cross-sectoral dispersion have decreased by more than two-thirds. This result may reflect the conviction that the latest lockdowns are, to some extent, expected to be the last, raising the likelihood of a gradual reopening of the economy.

Public link

Please use the above public link if you want to share this noodl on another website.