The astoundingly useful guys at FT Alphaville pointed me towards this Gerard Minack analysis of emerging market returns yesterday.
The message is that high growth economies don’t necessarily translate to high equity returns.
The argument can be summarised as this:
- Earnings growth is correlated with economic growth
- Valuation changes contribute significantly to equity returns and can have a major impact on equity returns distinct from underlying economic growth for long periods, 10 or 20 years
- But in the long term these valuation changes should even out. We should still be left with a correlation between economic growth and equity returns
- High growth economies need significant investment. This additional investment in companies comes at the cost of equity dilution. High growth economies are positively correlated with high dilution.
- Thus, EPS correlation with economic growth is significantly lower than it would be without dilution.
- This explains the virtually zero correlation between dividends and economic growth
Check out the full story for some pretty graphs.
What’s interesting for me here is that none of these arguments require or allow for market efficiency. It’s a totally separate way of looking at the issue with empirical evidence to support it.
I suppose the market efficiency counter would be that the change in valuation over long periods should be exactly as required to provide an appropriate risk-adjusted return to investors given the expected changes in all other variables. I don’t know if I buy that or not.
The key message for me is the counter argument to the “obvious” view that high growth emerging markets necessarily provide greater equity returns in the long run. The same can be said for why high growth companies don’t necessarily provide higher equity returns in the long run. As the low-growth companies are spitting out dividends to investors, the high-growth companies are diluting existing investors as they raise more capital.
The one question I haven’t full settled in my own mind is whether real dividends being correlated with economic growth is the best measure. High dividends now should result in low dividends in future. Low dividends now should result in high dividends in future. We should expect a point-in-time correlation between high growth economies (and companies) and low dividend yields. I would think that this correlation is needed in addition to the time series analysis performed by Dimpson, Marsh and Staunton since there can be weird lag effects that diminish the correlation there.
All the same, food for thought, especially living in a low-moderate growth emerging market country!