The dividend swap market has been very active and has made strong gains this year. In this Learning Curve we consider European dividend swaps and estimate where we expect them to go from here. We find that the Euro STOXX 50, SMI and CAC dividend swaps still offer good upside while FTSE 100 and AEX dividend swaps are likely to stagnate or drift lower.
Introduction
In order to forecast the likely direction of the index dividend swap market, we make assumptions about who the potential investors are and what utility function they generally employ to make their investment decisions. Specifically, we assume that the investors in this product are mostly hedge funds and that they look to achieve an annual hurdle rate of 10-15%.
For each of the Euro STOXX 50, CAC 40, FTSE 100, SMI and AEX indices we compute bottom-up dividend forecasts using IBES consensus estimates. We then use these forecasts in conjunction with the current dividend swap levels to estimate the potential return. Lastly, we make some assumptions about the capital required to support these positions and hence estimate the annualised return on capital for each index dividend swap.
We also provide an analysis of these dividend swaps, whereby we discount each of the individual company dividends according to that company's credit spread and the risk free rate. This is to:
1) Highlight that the estimates are potentially subject to credit risk (indeed, dividends are generally subordinate to payments to creditors) and,
2) Attempt to quantify this risk to ascertain whether the value we find is not due to assumed credit risk.
Estimating Dividends Swap Returns
We first look at dividend swap investments from the perspective of expected profit and estimated capital required to support the position. Given the dividend swap market's primary investor base is the hedge fund community, we feel that their utility function (ie, hurdle rates) will be the primary determinant of buying/selling these products. In particular, we assume that the typical hedge fund requires a return on capital of between 10-15% for its investments and will use leverage to achieve this.
Using IBES Estimates To Forecast Returns
We first look at how much scope for upside there is from current dividend swap levels. As a reference point for dividends to be paid, we use IBES consensus dividend estimates.
We compute overall returns by taking the current dividend swap offers as the levels at which we buy and the IBES estimates as the level at which we would sell at expiry. We then annualize this return to compare performance across different maturities.
Assessing Capital Required
The returns (shown above in Figure 1) assume that investors pay this amount upfront and realise profits at expiry. In practice, dealers require an initial margin and then mark-to-market profits and losses from there. We now make two further assumptions:
1) Initial margins are approximately 5% of the notional.
2) In the worst case the mark-to-market loss will be approximately -15%.
Under these assumptions, investors would be able to leverage themselves by a factor of 5x1. Multiplying the return above by this leverage factor allows us to estimate the overall return on capital for each dividend swap investment. The Euro STOXX 50 clearly stands out as the best investment based on this analysis.
If we assume that investors will purchase dividends according to this utility function, then we can also use reverse engineering to obtain forecasts for dividend swap levels. That is, for each index we look for the highest dividend swap level at which investors would achieve exactly 15% returns (ie, the required hurdle rate), again assuming IBES estimates for future paid dividends.
Discounting Dividends For Credit Risk
One issue that is sometimes neglected in the context of dividends is the inherent credit risk. It is not clear what each analyst dividend estimate represents. That is, if analysts estimate dividends but already account for credit risk, then further discounting would be overly conservative. On the other hand, if analysts estimate dividends without accounting for any credit risk, then the cost of this dividend cash flow should be discounted. This introduces two further issues:
1) The appropriate discount rate--dividends would be discounted at a higher rate than bond cash flows.
2) Uncertainty around the dividend levels--they could be higher or lower.
Given that companies are generally reluctant to cut dividends, we assume that this uncertainty is biased toward the upside. Accordingly, we are slightly more comforted in assuming that these two points cancel each other out. In the following analysis of the Euro STOXX 50, we discount IBES expected dividend levels by a discount rate composed of the credit spread and risk free rate.
We first decompose the Euro STOXX 50 dividend swap into its fixed and floating legs. The fixed leg is known and does not involve the credit risk of the underlying companies, so we discount cash flow by the risk free rate. On the other hand, the floating leg is subject to the credit risk of the underlying companies. We further break down the index floating leg into individual payments by each of the 50 companies and discount each by a discount rate equal to the risk free rate plus the company's relevant CDS spread.
Figure 4 shows the results of this analysis. Most importantly, it appears that the Euro STOXX 50 dividend swaps appear to offer value even with (potentially overly conservative) credit discounting.
Other Factors Affecting Dividend Outlook
While the Euro STOXX 50 looks to be the most attractive dividend swap investment based on this analysis, other factors could alter the outlook for the dividends paid by these indices.
In France, the recent abolition of the avoir fiscal benefit (2004 French Finance Act) for shareholders could cause investors to push companies for higher payouts. We feel that this provides an upward bias for future French dividend distributions.
The payout ratio and absolute level of the dividend yield will also likely be a factor for the future direction of dividend distributions. In particular, companies that already pay high ratios and/or yields could be unable or unlikely to increase their payout, whereas the opposite could be true for those with low payout ratios and yield.
Among the indices we are looking at, the SMI and AEX indices are the lowest and highest yielding, respectively. Moreover, the SMI has the lowest payout ratio (30%) while the AEX payout (59%) is second highest. The FTSE 100 has the highest payout ratio (74%) and second highest yield. Moreover, new pensions regulation could provide an obstacle to dividend growth as the FTSE 100 collective pension deficit is approximately 3% of total market capitalization. The Euro STOXX 50 falls in the middle in both the yield and payout categories.
Lastly, there are some cases where IBES estimates underestimate what the market is pricing. For example, IBES still estimates Total will pay just over EUR6 in 2006 whereas the market is pricing in more than EUR7. Moreover, Munich Re just announced a substantial increase in their dividend and analysts will likely be revising their estimates accordingly.
Conclusions
Taking these points and the IBES estimates analysis into account, we feel that the Euro STOXX 50 dividend swaps offer the most upside potential from current levels. We would also argue that the SMI and CAC dividend swap levels offer good value as well. On the other hand, we find that FTSE 100 dividend swaps offer the least upside and face the most obstacles. The AEX dividend swaps appear neutrally priced based on IBES estimates but we are wary for the scope for further upside given its already high yield and payout.
We suggest investors go long Euro STOXX 50 and CAC dividend swaps in 2008-2010, and SMI dividend swaps in 2008 (or longer maturities once they become liquid). For investors looking for short or relative value plays, we suggest selling the FTSE 100 dividend swaps in 2008-2010.
This week's Learning Curve was written by Aaron Brask and Priya Balasubramanian, equity derivative analysts at Barclays Capital in London.
