A perpetual growth DDM approach suggests an unrealistic amount of growth is needed to justify investment in US (versus European) stocks and in basic resources and construction (in both regions)….
Sign up for our weekly Newsletter and receive the latest ETF and ETC news. Click here to register for your free copy
Most investors make decisions based on projected returns. Hence, they are implicitly making a forecast about the future (how will central banks behave, how will bond yields change, what multiple will the market apply to the projected stream of earnings and dividends etc.).
This approach has one major drawback: we are all pretty bad at forecasting. An alternative approach incorporated by Andras into his Sector Selector document is to do the reverse. Starting with what we consider to be an acceptable rate of return (the hurdle rate of return), we ask what would need to happen in the future to generate those returns and then decide if that is plausible.
A simple perpetual growth dividend discount approach enables us to make judgements about US and European (including UK) equity markets, as well as their component sectors. Figure 1 shows the results based on data as at the end of 2017 Q1.
The implied growth rates are real, thus avoiding the distorting influence of inflation; the higher the growth rate the more optimistic one needs to be to justify investment. Our analysis suggests more growth is needed in the US than in Europe. Indeed, the implied perpetual growth rate for the US is 3%, which may be quite challenging given the reduced potential for economic growth (we think). On the other hand, the 0.2% real dividend growth implied for Europe seems quite conservative. On this basis, we believe that European equities are the more sensible investment.
When it comes to the sector detail, it is interesting that in both regions telecoms and utilities appear to be good value (the dividends can shrink every year forever and still give an acceptable rate of return). Basic resources is at the other end of the spectrum: an implausible amount of dividend growth seems to be required to justify investment.
Figure 1 – Implied perpetual real dividend growth rates (using actual bond yields, historical betas)
Note: based on a perpetual dividend growth model using the following inputs: current market based real bond yields: dividend yield using IBES consensus estimates of the next full year dividend payment, five year betas measured relative to local index (weekly data), risk premium based on US equity and bond returns since 1871. As of 31/3/17. Source: Datastream, IBES and Source Research
For instance, if dividends in the US basic resources sector were to grow at the implied 4.7% p.a., while the economy grew at less than 3%, those dividends would eventually be bigger than the economy! Likewise, the US construction and materials sector would need perpetual growth of 5.5% to justify purchase – that clearly cannot happen, no matter how many walls Donald Trump builds.
Before running out and implementing the above conclusions, we need to think about the methodology and consider how changes in the underlying assumptions would impact the results. As in any perpetual dividend growth model, growth is derived as the difference between the discount factor (risk-free rate plus beta-adjusted risk premium) and the dividend yield (g = r – d). For the risk-free rate, we use 10 year real sovereign yields: the TIPS yield in the US (0.42%) and the average of UK and Eurozone inflation protected yields (-1.5%). For the risk premium, we use the realised annualised excess return on US equities versus treasuries since 1871 (4.2%). We use this for Europe as well as the US, which is not ideal but probably as good as we can do given the absence of an equivalent long term series for Europe. Dividend yields are based on IBES consensus estimates of the next full year of dividend payments (to integrate market knowledge). Finally, we use sector betas as measured relative to the local market index over the last five years (using weekly data).
Choosing the appropriate risk premium is tricky. As mentioned above, we use actual US outcomes since 1871 as a proxy for what investors might consider the appropriate excess return on equities versus government bonds. If we used data from 1950, the implied risk premium would be 3.9%, rather than 4.2%, which arithmetically would reduce the perpetual growth rates needed to justify investment. On the other hand, data since 1980 would give a risk premium of 14.8%, implying much higher hurdle rates of return and therefore higher required growth rates. In order to abstract from those “short-term” differences, we prefer to stick with the longest historical time-frames possible.
Perhaps the most contentious input into our model is the current real bond yield (0.42% in the US; -1.5% in Europe). If we are simply concerned with beating the return offered by government bonds, that makes sense. However, those yields are extremely low and many equity investors will be more ambitious than that. Indeed, it could be argued that we should allow for a normalisation of bond yields when looking to the long term. Figure 2 shows the results if we use a real bond yield of 2% in the US and 1.5% in Europe (based on what we think might be reasonable estimates of long term economic growth). Not surprisingly, Figure 2 looks just like Figure 1, except the implied growth rates are higher (the ordering of the sectors is the same).
Figure 2 – Implied perpetual real dividend growth rates (using normalised bond yields, historical betas)
Note: based on a perpetual dividend growth model using the following inputs: “normalised” real bond yields (2% in the US, 1.5% in Europe): dividend yield using IBES consensus estimates of the next full year dividend payment, five year betas measured relative to local index (weekly data), risk premium based on US equity and bond returns since 1871. As of 31/3/17. Source: Datastream, IBES and Source Research
On this basis, our model suggests the perpetual real growth rates required to generate the hurdle rate of return is now 3.1% in Europe and 4.5% in the US, well above what we think is possible, given potential economic growth rates. Neither European nor US equity markets appear to be priced to offer what we might consider a “normal” rate of return.
Finally, in all that has gone before we have assumed that sector betas will continue to be as observed during the last five years. Not surprisingly, the sectors with the highest hurdle rates of growth in Figures 1 and 2 are also those with the highest beta. However, betas do change over time and it seems unlikely that sector betas can vary greatly from 1.0 over long periods of time (to see why, think about what would happen to a high beta sector if markets continue to rise over the infinite future).
Hence, in the final iteration, we continue to use “normalised” real bond yields (as in Figure 2) but now set all the sector betas to 1.0. Although the implied growth rate for the total market is not impacted by this process, it is clear from Figure 3 that the spread of sector implied growth rates is narrower (compare with Figure 2).
The telecoms and utilities sectors continue to feature among those that require the least growth to generate acceptable returns but the margins are now finer. Autos, oil & gas and financials are examples of sectors that look more attractive under this methodology. At the other end of the spectrum, the basic resources and construction sectors remain among those needing to generate the highest rates of growth to justify investment, though the comparisons are no longer so bad. Food & beverage, personal & household goods and healthcare are examples of low beta sectors penalised by this approach.
It has hopefully become apparent that the outcomes depend critically upon the assumptions. Lessons that we would draw are: whether equities are cheap or expensive depends upon the yardstick used (compared to existing bond yields, the hurdle is not too high but judged in absolute or “normalised” terms equities may struggle to produce satisfactory returns); the US corporate sector needs to generate much higher rates of dividend growth than does its European counterpart (this fits with our Overweight in Europe and Underweight in the US ); the dull and boring sectors of telecoms and utilities (where we are largely Underweight) can get away with subpar rates of growth and still give good returns (because the dividend yields are high); the more exciting basic resources and construction sectors need an unrealistic amount of growth to justify investment, which is a sobering thought as we like the short term dynamics of construction. Will Donald come to our rescue?