With reporting season drawing to a close, we look back at the historical accuracy of the market’s aggregate EPS forecasts made by sell-side analysts.

Sell-side analysts rarely underestimate EPS growth

Using aggregate estimate data from S&P Global, for CY2010* through to CY2024 we looked at how analyst full-year EPS estimates for the ASX200 and S&P500 changed each month. As is clear from Figure 1 and Figure 2, almost without exception, there is a downward trend, with forecasts progressively decreasing as the period end approaches.

On average, estimates for the ASX200 were 12% too high 2 years before period end, and 6% too high 1 year out. Excluding the two years where estimates were below actuals, the overshoot averaged 18% and 10% on a 2 and 1 year basis, respectfully.

Estimates for the S&P500 were significantly better at 10% and 1% on 2-year and 1-year basis, respectively. Excluding the under-estimated years the overshoot was 12% and 6% respectively.

‍

‍

Australian analysts more optimistic than US counterparts

Perhaps the most surprising result is that Australian sell-side analysts seem more optimistic than their American counterparts (or maybe the Australian market is more difficult to forecast). The percentage difference between forecasts and the actual EPS results for the ASX 200 was between 1.5 – 2 times the difference between forecast and actuals for the S&P500. While there is a lot of variation (see Figure 1 and Figure 2) between the years, a consistent trend across both markets is that sell-side analysts are consistently optimistic in aggregate.

‍

‍

For the ASX200, there were only two years (out of 14) where the year finished with estimates above the 12-month forecasts and these were during COVID (CY21 and CY22). The 24-month out forecasts were similar, with only the results for CY22 and CY23 ahead of expectations.

US analysts were only marginally better when measured by directional calls. Two years out, the analysts were overly optimistic 13 of 15 years, but one year out, it was slightly more mixed with four years surprising on the upside (see Table 1).

While Australian and US analysts made similar mistakes directionally, the magnitudes of the over-estimations are meaningfully different. As shown in Table 1, the forecasts of the ASX200 24 months before the end of the period were 1.5 – 2 times more erroneous compared to S&P500 estimates. The gap narrowed when looking at the 1 year out estimates.

‍

‍

Several factors impact the accuracy of market forecasts at the aggregate level, including:

1) Most US companies provide quarterly results

2) Management of US companies tend to provide more guidance

3) Due to the market size, US analysts can specialise in sub-sectors while Australian analysts tend to cover multiple sectors.

4) While the S&P500 is more concentrated at the very top (23 companies represent 50% of the market cap vs. 13 for the ASX200), it has a marginally stronger tail with 239 companies representing 90% of index market cap vs 85 for the ASX200). As such, outside the top 10-20 companies, missed estimates of individual companies in the S&P500 will have less of an impact.

Ultimately, on aggregate, based on the historical data, sell-side analysts are consistently over-optimistic when estimating companies’ financial results.

‍
* ASX200 estimate data started in Feb-2010 so thefirst 12 month comparison year is CY11, and the first 24 month comparison yearwas CY12