I may be wrong, but there seems to be a misunderstanding. My understanding is that aggregate earnings growth is usually - to some extent - reflected in expected per share earnings, minus a correction factor. The studies that show no or negative correlation between growth and returns are based on actual historic returns, with actual valuations at the start and end of the evaluation period. Countries with higher expected returns usually have higher valuations, all other things being equal. The market usually overvalues growth. As a result, countries with higher growth usually have lower returns. However, if valuations are the same, higher growth countries are expected to have higher returns based on most investment banks' models. So the conclusions appear different because of different assumptions on which parameters are held constant.
I did not study how this reconciles with the numbers presented earlier in this thread. But a more sophisticated analysis can be found in many "global outlooks" by the major investment banks. https://www.blackrock.com/institutions/ ... ssumptions
shows an example of a model. "Components of Five-Year Requity Return Assumptions", they have quite different numbers for "net share dilution" for US and DM Ex-US. I think I saw bar diagrams showing the return components for US, DM, EM on this site a few weeks ago, but I can't find them any more. Coincidentally, contrary to the common belief that the U.S. is hugely overvalued compared to international, their US and DM total return forecasts are almost the same, after backing out Re-Pricing and Adustments. I have a hard time understanding the rationale for their "adjustments for the different risk-free cash returns between local-currency cash and US cash". How does the cash rate affect returns?)
https://am.jpmorgan.com/blob-gim/protec ... cale=en_US
(click "Advisor"; email registration may be required to download) , page 63, shows earnings growth, gross dilution, and buyback assumptions for various DM and EM countries.
Regardless of actual numbers, most models agree that EM have higher expected returns based on current valuations, as growth is higher AND valuations are lower (although slippage / dilution is also higher).