IL Securities and Inflation Risk Management NYC, February 10th, 2009

Transcription

1 IL Securiies and Inflaion Risk Managemen NYC, February 10h, 2009 Inflaion-Hedging Properies of Real Asses and Implicaions for Asse-Liabiliy Managemen Decisions Lionel Marellini Professor of Finance, EDHEC Business School Scienific Direcor, EDHEC Risk and Asse Managemen Research Cener

2 Ouline Moives for inflaion hedging Insrumens for inflaion hedging Inflaion hedging wihou inflaion-linked asses From shor-erm o long-erm inflaion hedging Implicaion for pension funds

3 Moives for inflaion hedging Insrumens for inflaion hedging Inflaion hedging wihou inflaion-linked asses From shor-erm o long-erm inflaion hedging Implicaion for pension funds

4 Inflaion Hedging: Why? Key insigh of dynamic asse pricing heory is he inroducion of dedicaed hedging demands (in addiion o sandard performance-seeking porfolio), allowing invesors o hedge away impac of unexpeced changes in relevan sae variables (Meron (1969, 1971)). Inflaion is one key sae variable ha opimally commands an hedging demand for he following possible reasons: Impac on wealh process (endowmen and liabiliy sreams); Impac on produciviy of wealh (sochasic opporuniy se); Impac on uiliy of wealh (choice of a numeraire). Recen papers on inflaion in dynamic asse allocaion models: Brennan and Xia (2002), Campbell Chan and Viceira (2003), Sangvinasos and Wacher (2005), Hoevenaars e al. (2008), Deemple and Rindisbacher (2008), Marellini and Milhau (2008), among ohers.

5 Recen Increase in Inflaion Volailiy

6 Inflaion: Dead bu no Buried Curren credi crunch and relaed economic slowdown has eased he 2008 inflaion scare for he nex year, bu inflaion concern remains vivid for he medium- o long-erm. The impac of he emerging economies on prices has unambiguously urned from being deflaionary o inflaionary, and his rend is bound o las. Need for governmens o suppor weak economies wih significan simulus packages will remain, along wih injecion of liquidiies in money markes.

7 Moives for inflaion hedging Insrumens for inflaion hedging Inflaion hedging wihou inflaion-linked asses From shor-erm o long-erm inflaion hedging Implicaion for pension funds

8 Inflaion Hedging: How? A key quesion from he empirical perspecive in an incomplee marke seing is he opimal composiion of inflaion-hedging porfolios. Inflaion-linked securiies, and mos imporanly inflaionlinked bonds, are a priori he mos naural hedge agains inflaion. Governmens (and mos corporaions) are naurally hedged agains inflaion on he asse side, and find i opimal o issue a significan amoun of IL bonds ha allow hem o benefi from inflaion risk premium wihou increase in (ALM) risk (Marellini and Milhau (2008)). Inflaion swaps are also used o cusomize inflaion exposure.

9 Inflaion Hedging: How? Ousanding quesions: Can invesors achieve perfec hedging? Do invesors really need hedging a each dae? Incomplee markes: Supply side: capaciy of IL securiy marke no sufficien o mee he collecive demand of insiuional invesors. Demand side: pension paymens are ofen indexed wih respec o wage, as opposed o price, inflaion. Long-erm invesors: Problem in ALM is no hedging inflaion dae by dae, bu insead hedging inflaion a a ypically long horizon. Shor-erm hedging (when possible) can evenually prove unnecessarily cosly.

10 From Price o Wage Inflaion CPI is he Consumer Price Index For All Urban Consumers and WAS he index eniled Compensaion of Employees: Wages & Salary Accruals. Boh series are seasonal adjused. Deails on hp://research.slouisfed.org/fred2/.

11 Moives for inflaion hedging Insrumens for inflaion hedging Inflaion hedging wihou inflaion-linked asses From shor-erm o long-erm inflaion hedging Implicaion for pension funds

12 Inflaion Hedging Properies of Socks Socks On he one hand, equiy invesmens appear as relaively poor inflaion hedging vehicles from a shor-erm perspecive. Empirical evidence indeed suggess ha here is a negaive relaionship beween expeced sock reurns and expeced inflaion, consisen wih he inuiion ha higher inflaion leads o lower economic aciviy, hus depressing sock reurns (e.g., Fama and Schwer (1977)). On he oher hand, higher fuure inflaion leads o higher dividends and hus higher reurns on socks, and hus equiy invesmens should offer significan inflaion proecion over longer horizons (Campbell and Shiller (1988)), a fac ha has been confirmed by a number of recen empirical academic sudies (e.g., Boudoukh and Richardson (1993)).

13 Inflaion Hedging Properies of Bonds Bonds Bond yields may be decomposed ino a real yield and an expeced inflaion componens. Since expeced and realized inflaion move ogeher on he long-erm (Schoman and Schweizer (2000)), we expec a posiive long-erm correlaion beween bond reurns and changes in inflaion. In he shor-erm, however, expeced inflaion may deviae from he acual realized inflaion, leading o low or negaive correlaions (rigidiies of he nominal yield due moneary policy may cause a rearded response).

14 Inflaion Hedging Properies of Commodiies Commodiies Commodiy prices generally are se in highly compeiive aucion markes and consequenly end o be more flexible han prices overall. Beside, recen inflaion is heavily driven by he increase in commodiy prices, in paricular in he domain of agriculure, minerals and energy. Paricularly well suied in he shor run as idiosyncraic commodiy price shocks induce inflaion (naural caasrophes). Goron and Rouwenhors (2006) find ha, over he period, commodiy fuures were posiively correlaed wih inflaion, unexpeced inflaion, and changes in expeced inflaion.

15 Inflaion Hedging Properies of Real Esae Real esae Real esae income should reflec pas and curren inflaion. Real esae price level is par of he CPI baske inducing an endogenous link beween real esae prices and CPI inflaion. However, rens and real esae prices exhibi some rigidiies wih respec o curren inflaion shocks. Empirical evidence of inflaion-hedging benefis of real esae invesmens can be found in Fama and Schwer (1977), Harzell e al. (1987) or Rubens e al. (1989), Hoesli e al. (2007).

16 Moives for inflaion hedging Insrumens for inflaion hedging Inflaion hedging wihou inflaion-linked asses From shor-erm o long-erm inflaion hedging Implicaion for pension funds

17 Economeric Mehodology Evidence for reurn predicabiliy (e.g. Kandel & Sambaugh (1996), Campbell & Viceira (1999)) based on vecorauoregressive (VAR) models capuring mean-reversion effecs. Mean-reversion in sock reurns is induced by posiive relaionship (β 2 >0) wih mean-revering variable (DY). Price levels and heir firs differences ener he above model. + + = p d rf r c c p d rf r 2, 1, ε ε β α β α r R P P P P p = + = = = ) log(1 log ) log( ) log( 1 1

18 Economeric Mehodology Vecor Error Correcion Model (VECM): add informaion on long-erm price equilibrium relaionships in addiion o pure reurn modelling allows one us o explicily disinguish beween shor-run and long-run dynamics (wih meanreversion owards long-erm price relaionship β). y + 1 = c + αβ ' y + Γ y + ε +1 The model involves insananeous shocks o he sysem if i deviaes from he long-erm equilibrium αβ. VAR and VEC models deliver model-implied forward-looking erm srucures of risk and reurn parameers via he process of innovaions ε.

19 Economeric Mehodology Tradiional asses: Socks: CRSP vw index of NYSE and AMEX lised socks Bonds: Lehman US Treasury Aggregae T-Bill: 1 monh reasury bills from CRSP Alernaive asses: Commodiies: Goldman Sachs Commodiy Index (GSCI) Real esae: FTSE NAREIT real esae index Economic variables: CPI index Credi spread: difference of Baa and Aaa Corporae Bond yields Term spread: difference beween 10Y and 3M Treasury yield Dividend yield from CRSP Daa from Q hrough Q

20 Term Srucure of Volailiies

21 TS of Correlaions w.r.. o Liabiliies

22 Moives for inflaion hedging Insrumens for inflaion hedging Inflaion hedging wihou inflaion-linked asses From shor-erm o long-erm inflaion hedging Implicaion for pension funds

23 Risk Budgeing in ALM In erms of risk budges, he implemenaion of ALM/LDI soluions criically depends on he aiude owards risk. High risk aversion levels leads o a predominan invesmen in he LHP, which in urn implies low exreme funding risk, low expeced performance and high necessary conribuions. On he oher hand, low risk aversion levels lead o a predominan invesmen in he PSP, which implies high funding risk, higher expeced performance, and lower conribuions. Compare he iniial conribuion needed o generae a 100% funding raio a he horizon when he invesor's porfolio is fully invesed in TIPS (he perfec LHP) versus he iniial conribuion needed o generae an average 100% funding raio a he horizon when sock-bond PSP is inroduced.

24 Inroducing a PSP If horizon = 20 years, an allocaion of 40% o he PSP => reducion of he iniial conribuion by almos 20%; of course, his comes a he cos of inroducing funding risk. 40% PSP Conribuion savings defined as he reducion of iniial invesmen when inroducing a PSP so ha, on average, he pension plan is fully funded. PSP conains socks and bonds in proporions ha depend on horizon.

25 Enhancing he LHP

26 Enhancing he LHP If horizon = 20 years and LHP is enhanced (10% AI), only need 28% PSP o reduce conribuions by same amoun as when using 40% PSP, and 60% TIPS LHP. 20 Y This figure shows he impac of enhancing he LHP beyond TIPS. The enhanced LHP consiss of 90% TIPS and 10% alernaive invesmens. The graph shows he required allocaions o he PSP needed o obain on average he same expeced reurn as wih he LHP (TIPS).

27 Risk Reducion If horizon = 20 years and LHP is enhanced (10% AI), can reduce he allocaion o PSP by 30% (28% insead of 40%), which leads o an 40% decrease in probabiliy of shorfall and almos 80% decrease in probabiliy of an exreme shorfall. We consider he base case of an invesmen of 40% in he PSP and 60% in he radiional LHP and compare his porfolio o he case where he LHP is enhanced by 10% AI.

28 Conclusion Some alernaive forms of invesmen (in paricular commodiies and real esae) exhibi aracive inflaion hedging properies while offering compelling reurns. Inroducion of suiably seleced alernaives in insiuional invesors porfolios allows for a significan reducion in he required allocaion o he PSP for a given level of performance (or required conribuions). As a resul of inroducing liabiliy-hedging porfolios wih enhanced performance, as opposed o generaing performance hrough he performance-seeking porfolio only, funding risk is significanly reduced. PSP could also be improved by he inroducion of alernaive asses.

29 Exensions While we have focused on real esae and commodiies, insiuional invesors have recenly shown an increasing ineres in oher alernaives such as privae equiy and infrasrucures, for which he inuiion suggess ha aracive inflaion-hedging properies could also be obained. More generally, one may envision o selec securiies or asse classes on he basis of heir liabiliy hedging properies, as opposed o selecing hem on he basis of heir expeced abnormal performance. On he alernaive side, precious meals and reail real esae, among ohers, show paricularly aracive liabiliy LT hedging benefis.