Abridged from ABS Dilemma, by Tom Zimmerman, ABS analyst, PaineWebberAs Treasury market de-linkage becomes more apparent, alternative sectors have become more important as benchmarks. In particular, the swap market grew in importance to the ABS market. Emergence as a key force behind ABS spreads did not develop quickly - it evolved over some years. Several years ago in the ABS market, cards and autos were closely linked to movements in swaps; the mortgage-related sectors (such as home equities) had only a loose linkage. A large movement in swap spreads would eventually impact home equities, but not as fast as it registered in the card or auto sectors.
However during the past two years, the growing awareness that outstanding Treasury supply would be shrinking (and possibly disappear altogether) increased the market's use of swaps for hedging and benchmarking in all parts of the ABS market, including mortgage-related sectors. Swap movements now have a very distinct impact on home equities, even though that particular linkage is still not as immediate as in the card and auto sectors.
Swap Market Volatility
One of the problems with benchmarking via swaps is the complexity of forces determining swap levels. Figure 1 shows swap spreads since January 1999. Swap spreads, and credit spreads in general, were relatively stable until May. Then they and other spread product widened dramatically. All peaked in July and August, as supply imbalances and Y2K issues (remember those?) increased credit and liquidity concerns. By mid-January the easing of concerns post-Y2K took swap spreads and most ABS spreads back to or below their early 1999 levels. However, Treasury's announcement that they would be more aggressive in buying back their debt triggered a marked inversion of the Treasury curve (beginning in late January). And with that, longer-dated swaps took off.
The recent flap about Treasury backing a bill redefining implied government support for Agency securities caused a further widening of long-dated swaps. The 10-year widened to 125, well above its prior record of 110 from last summer (and above the level reached in October 1998), and far above the end of January's 70.
Figure 2 shows the difference in spread between 2- and 5-year swaps, which has just surpassed the highs of last August. In contrast, the 5- and 10-year swaps differed by 10-15 basis points most of last year, and now hover around 30 basis points apart.
So we now have several powerful forces causing a great deal of uncertainty. An inverted Treasury yield curve with no prospect of normalizing in the foreseeable future; a growing quandary about what benchmark, if any, to replace Treasuries; and a volatile swap market.
When the 10/30s segment of the yield curve first inverted it did not impact the ABS market dramatically, since most ABS securities are either floaters or are relatively short fixed rate. It was the long corporates that bore the brunt of the first inversion in late January. However shortly thereafter the 5s/10s inverted, and 2s/5s have now also inverted. Clearly that inversion, the diminished role for Treasuries, and swap volatility - are all impacting ABS spreads.
Lessons From Recent Pricing
To illustrate the convoluted forces now at work in the ABS market, we examined changes in home equity spreads over the past five months versus what happened in concurrent Treasury and swap markets.
Table 1 shows pricing of several home equity deals in terms of spreads for 2-, 3-, 5- and 7-year classes. All issuers are Tier 1, and the deals trade fairly close to one another. Spread changes over this period varied a great deal, depending on maturity. The 2-year changed little, the 3-year widened a few basis points, the 5-year widened 10+, while the 7-year was 25 wider. Shown next to the nominal spreads are corresponding swap spreads, and then on the far right are the home equity spreads swapped out.
In Table 2 we show the difference in spreads between adjacent home equity average life classes (from Table 1). Next to that we list the differences between Table 1's swap spreads. Next to those we show the Treasury curve for 2s/5s and 5s/10s for the same dates on which the home equity deals were priced. We used a three-day moving average to dampen daily yield volatility.
We can now examine changes in both the Treasury curve and the swap curve to see which provides a better explanation of the changes in home equity spreads, and whether any relative value conclusions can be drawn.
First look at the spread differences in Table 2 for home equity maturities of 2- and 5-years. The difference increases from 30, to 44, basis points over this period, for a net increase of 14 basis points. Now look at the 2s/5s part of the Treasury curve shown in Table 2. That part flattened from around 14 basis points to only 2 basis points by 3/14/00 (and is now inverted by 13 basis points), for a change of around 12 basis points. This is almost the same as the change in home equity spreads. It appears that as the five-year Treasury richened vs. the two-year Treasury, the 5-year home equity widened vs. the five-year Treasury by a like amount - keeping the yield between 2- and 5-year home equities pretty much where they were relative to each other in November.
Let's now look at swap spreads during this same period. As shown in Table 2, the difference in spreads on 2-and 5-year swaps changed very little. In November and December last year, it was 15 and 18 basis points, respectively. On the last two observation dates (3/9/00 and 3/14/00), it measured 15 and 19. Yes, both 2- and 5-year swaps moved out sharply in the interim, but the difference between them was little, if at all, changed.
Here is a case where home equity spread relationships were more closely linked to changes in Treasuries than to changes in the swap market. Because both swap spreads widened roughly the same amount, while home equity spreads did not (2-years were about unchanged, while 5-years widened 12 basis points), the relative value from a swapped perspective also changed. As shown in the right section of Table 1, 2-year swapped home equities look pretty rich compared to where they were trading in November, while 5-year swapped home equities look about the same.
So the question is, "Are 5-year home equities richer or cheaper versus 2-year home equities than they were in November?" Versus the swap market the 5-years appear cheaper. Yet in simple yield terms there has not been much of a change.
A Longer Look
To further explore the relationship between shifts in the Treasury curve and the swap market, and changes in ABS credit curves, we conducted the same exercise as before, only with generic ABS spread data over a longer period. This time, we looked at credit curve changes for both credit cards and home equities going back to June 1999 (prior to the July/Aug spread widening and long before hints of a serious yield inversion).
Consider credit card data. Five-year card spreads were 20-23 basis points wider than 2-year spreads back in June and July of 1999. That difference contracted sharply by November, and by late January (just before the inversion and the swap market explosion) it hit a tight of 10 basis points. By March 17, the 5-year was again 23 basis points wider than the 2-year. Essentially there was no change over the period in the difference between spreads on the 2- and 5-year cards.
Meanwhile, the 2s/5s part of the Treasury curve went from a positive slope of 18 basis points to a -4 basis point inversion, for a net change of -22 basis points. Since the difference in spread between 2- an 5-year cards was about unchanged, this meant the yield on the 5-year card relative to the 2-year dropped by around 22 basis points.
If we look at the swap spreads we see a very different picture. The change in 2/5 swap spreads (the difference between 2- and 5- year swap spreads) looks very similar to the change in card spreads. In fact, the trend is almost the same. Not only is the change similar, but the absolute magnitude of the difference between 2- and 5-year swaps is almost the same as that between spreads on 2- and 5-year cards.
In this example card spreads were clearly influenced to a great extent by changes in the swaps market. The change in the shape of the curve was not a major influence.
Now look at the table's new home equity series, which covers a longer period than the data from Table 1 and uses generic rather than deal-specific spread data. This time, the data seems to support the view that swap spreads are the most important factor impacting home equity spreads. In June/July 1999, 5-year home equities were around 40 wider than 2-year home equities. That difference subsequently declined to 25-30 basis points, but has now returned to the 40+ level.
Meanwhile, the 2s/5s part of the Treasury curve has inverted. This leaves yields on the 5-year home equity much lower relative to the yield on the 2-year home equity than they were last June. However, on a swapped basis they are about the same relative level. This picture is different from that discussed earlier for home equities (based on data in Tables 1 and 2). In part, the difference results from taking observations over a longer time period. It also arises because of the specific dates used. Spreads and yields in all of these markets are volatile, and a market reading can change a great deal just by shifting the dates a few days.
No Clear Winner
The result of our exercise was inconclusive. It appears that parts of the ABS market strongly follow swap market leads. But sometimes, a sharp change in the shape of the Treasury curve can exert an even greater influence. Also, the time period over which one makes an analysis can deliver very different relationships - as can the specific dates chosen for the analysis.
Our advice? Beware of quick relative value comments in these volatile markets. The usual standards of spread differentials between different maturities no longer hold. Spreads are now quite volatile, and should only be viewed in the context of changes in both the shape of the Treasury curve and levels in the swap market.