EFTA01458271

I September 2015 Corporate Credit: Back to school - The edge of normality Figure 5: Average BBB Spread Change in the lead up to and post the First US Hike (left) and Last US Hike (right) in a Cycle. Data across 12 rate cycles since 1950. •BBB Credit Spread (Avg) •BBB Credit Spread (Med) 40 30 • 0 2 • 10 • -10 • -20 - -30 - -40 - -50 I glif irs trill ir e \c1.4% Act I, Kt V-6 41/41 stg itLe /tee er tr Sotto. Deatache Boni. GOV. Moons •BB8 Credit Spread (Avg) 70 60 50 40 30 20 100 -1 -20 • 888 Credit Spread (Med) cri ,—.0.-disbrA..a iii it ril IA es elite, g1/44} .t4-6 (Le ct* 4400 fr fr i CSL niti,bre./ 1•4‘ esereene The caveat would be that we haven't seen the spread tightening in this cycle in the lead-up to a potential hiking cycle that we normally do. This perhaps illustrates that the historical context for this cycle is highly uncertain but at face value the immediate impact of a rate hike is not usually negative for credit. This lag impact of a monetary hiking cycle fits in with our model for predicting the default cycle outlined in our 2015 Default Study published back in April. Here we showed how the level of real yields and the shape of the yield curve tended to help predict where the default rate would be in 30 months time. At the moment, real yields are still structurally low and although the US yield curve (between Fed Funds and 10 year Treasuries) has flattened from its most recent peak at the end of 2013, it is still relatively steep. The model (Figure 6) does forecast that perhaps the low point in defaults is behind us but we are still likely to stay below average over the next couple of years. The risks will start to materialise if the Fed starts a cycle of rate rises and the long-end rallies ensuring a significantly flatter yield curve. At that point our model would suggest a faster pace of rising defaults 30 months later. Figure 6 1.15 HY Default Rate Model vs. Actual Default Rate 16% Default Rate I%) —Model Default Rate I%) 14% R'=46% 12% 10% 8% 6% 4% 2% 0% 1988 1992 1996 2000 2004 2008 2012 2016 Savor Datchs Berk AbeWs Given that spreads tend to move around 9 months (on average) before defaults pick up (Figure 7) then maybe we can see how a rate hike cycle could start to materially hit credit markets just over a year and half after the first hike. Clearly this assumes old relationships hold which is clearly not a given in this unique cycle. However it's the only framework we have. Page 4 Deutsche Bank AG/London CONFIDENTIAL — PURSUANT TO FED. R. CRIM. P. 6(e) CONFIDENTIAL SDNY_GM_00264289 DB-SDNY-0118105 EFTA01458271