Why are real interest rates so low? M. Marx, B. Mojon, F. Velde Warsaw 17 December, 2015
Motivation (1/2) Why are interest rates so low? What can we do about it?
Motivation (2/2) The debate on the level of real interest rates is essentially qualitative. Poor prospect on economic growth (Gordon, Summers) 1 demographics lower investment and lower productivity 2 demographics higher savings and higher demand for safe assets, inc. by EM (IMF view) Debt deleveraging (which is somehow equivalent) (Koo, Eggertsson, Krugman) Drop in the supply of safe assets (Caballero and Farhi) We provide one attempt of a quantitative assessment of these various forces
Related Literature Eggertsson and Mehrotra (2014) Bianchi and Mendoza (2015) Buera and Nicolini (2014) Hamilton et al. (2015)
Outline 1 Interest rates are indeed low 2 Show where it can come from in the simplest possible model 3 Where in the data do we see anything like what the model is asking for
Evidence on real interest rates 8,00 7,00 6,00 EU SWAP 1Y US SWAP 1Y JAPAN SWAP 1Y 5,00 4,00 3,00 2,00 1,00 0,00
Evidence on real interest rates 8,00 7,00 6,00 EU SWAP 5Y US SWAP 5Y JAPAN SWAP 5Y 5,00 4,00 3,00 2,00 1,00 0,00
Evidence on real interest rates 4,00 3,00 2,00 1,00 0,00-1,00 EU_1Y US_1Y JP_1Y EU_5Y US_5Y JP_5Y -2,00-3,00 2004 2005 2006 2007 2008 2009 2010 2010 2011 2012 2013 2014 2015-4,00
Evidence on real interest rates 7 6 5 4 real_ea_1y real_us_1y real_jp_1y 3 2 1 0-1 -2-3 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Evidence on real interest rates 4,00 3,00 2,00 real_ea_5y real_us_5y real_jp_5y 1,00 0,00-1,00-2,00-3,00 2004 2005 2006 2007 2008 2009 2010 2010 2011 2012 2013 2014 2015
Evidence on real interest rates
Evidence on real interest rates 7 6 FR JP US 5 4 3 % 2 1 0 1 2 3 1985 1990 1995 2000 2005 2010 2015 real 1 year interest rates
Evidence on real interest rates :take away Real interest rates have moved from 2% to somewhere between 0 and -2% We care about -2% because it is their physical minimum with anchored inflation expectations at 2% Real interest rates move together (Hamilton et al, 2015)
Model adapted from Farhi and Werning (2013) a 2-country minimalist New Keynesian model 3 periods t {0, 1, 2}, initial, intermediate, final (5 years) 2 countries with relative population shares φ 1 and φ 2. Country 1 : linear technology to transform labor into output At embeds productivity and size of the working population Country 2 : only endowment E t Country 1 : savers/germany, Country 2 : borrowers/spain
Preferences Agents in country 1 work and consume in every period : V 1 s = 2 β t [ u(ct 1 ) v(nt 1 ) ] t=0 while agents in country 2 consume but do not work : V 2 s = 2 β t u(ct 2 ) t=0
Frictions 1 maximum rigidity : prices set once and for all 2 borrowing constraint on agent 2 in middle period ( deleveraging shock )
Basic mechanics the borrowing constraint reduces the demand for loans in the middle period if large enough, it drives down the nominal interest rate to 0 then, monetary policy can t stabilize fully macroprudential policy : a borrowing constraint on agent 2 in the first period can help under what parameter configuration will the ZLB bind in the middle period?
Ramsey approach to macro-prudential policy choose allocation {ct, i N t } to maximize total welfare subject to equilibrium constraints the constraints boil down to feasibility at each t, the ZLB, and deleveraging the ZLB u (ct 1 ) βu (ct+1 1 ) embodies all the price rigidities (constant prices) the social planner can do lump-sum transfers countries BC are not constraints this makes decentralized interpretation awkward at times
Autarky allocations are static : c 2 t = E t and c 1 t solves u (c 1 t ) = v (c 1 t /A t ) A t
Flexible price (no ZLB) Again, allocation is static and solves or c 1 t solves u (c 1 t ) = v (c 1 t /A t ) A t λ 1 u (c 1 t ) = λ 2 u (c 2 t ) φ 1 c 1 t + φ 2 c 2 t = φ 1 A t N t + φ 2 E t F (c 1 t, E t, A t ) = 0 F defines an increasing function C(E t, A t )
Calibration λ 2 = λ 1 relative shares φ 1 = 1, and φ 2 = 0.5 u(c) = c1 σ n1+η 1 σ, and v(n) = 1+η σ = 2, and η = 0.5 discount factor (1/1.02) N
Hitting the ZLB (representative agent) the ZLB condition for country 1 : ( A2 ) 1+η σ+η 1 < β σ or A < σ + η ( ) 1 A 1 σ(1 + η) β 1 = 2 ( ) 1 3 β 1 ( ) ( ) the ZLB condition for country 2 : E < 1 1 σ β 1 = 1 1 2 β 1
Necessary conditions for ZLB to bind Three ways to get the ZLB to bind : 1 There is a β-shock β S on the discount rate of country 1, such that at least, β s β > ( ) c(e2, A 2 ) σ c(e 1, A 1 ) 2 either E or A is decreasing (E 2 < E 1 or A 2 < A 1 ) and c(e 2, A 2 ) < β 1/σ c(e 1, A 1 ) 3 For a borrowing constraint on country 2 to bind, A (1+η)/(σ+η) 2 < min{(φ 2 /φ 1 ) 1/σ E 2, β 1/σ c(e 1, A 1 )}
Hitting the ZLB : with the discount factor 1.5 1.4 1.5 1.4 r=-2 r=0 r=2 1.3 1.3 β 1.2 1.1 1 0.9 β 1.2 1.1 1 0.8 0.9 0.7 0.8 0.6 0.7 0.8 0.9 1 1.1 1.2 0.8 0.9 1 1.1 1.2 A 2 E 2 Value of the discount factor ensuring the ZLB, E 1 = 1 and A 1 = 1
Hitting the ZLB : with the discount factor 1.6 1.5 1.4 1.3 1.2 1.5 1.4 1.3 1.2 r=-2 r=0 r=2 β 1.1 β 1.1 1 0.9 0.8 0.7 1 0.9 0.8 0.6 0.7 0.8 0.9 1 1.1 1.2 0.8 0.9 1 1.1 1.2 A 2 E 2 t = 10 years
Hitting the ZLB : with productivity and endowments 1.1 1.05 r=-2 r=0 r=2 1 0.95 A2 0.9 0.85 0.8 0.75 0.75 0.8 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 E 2 Impact on productivity and endowment, E 1 = 1 and A 1 = 1
Hitting the ZLB : with productivity and endowments 1.05 1 r=-2 r=0 r=2 0.95 0.9 0.85 A2 0.8 0.75 0.7 0.65 0.75 0.8 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 E 2 t = 10 years
Hitting the ZLB : with productivity and endowments 1.1 1 σ = 0.5 σ = 2 σ = 4 0.9 0.8 A2 0.7 0.6 0.5 0.4 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 E 2 changing σ doesn t make a difference
Hitting the ZLB : with borrowing constraint 1.05 r=-2 r=0 r=2 1 0.95 A2 0.9 0.85 0.8 0.75 0.8 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 E 2 t=5 years
Hitting the ZLB : with borrowing constraint 1.05 1 r=-2 r=0 r=2 0.95 0.9 A2 0.85 0.8 0.75 0.7 0.65 0.75 0.8 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 E 2 t=10 years
What features of the data look like what the model requires? Nothing!
Future US GDP/Income FOMC : Growth of potential GDP =2.75% per capita = 2% Gordon : Growth of GDP per capita (yoy growth) 1891-2007 2% 2007-2032 0.9% Mainly due to lower labor market participation 1972-1996 +0.4% 2004-2014 -0.5% 2007-2014 -0.8% Income distribution : yet not worse than since 1984 Public debt : Ambiguous
Future GDP/Income (OECD)
Debt deleveraging By how much has debt decreased /is required to decrease Actually, Global Debt has kept on growing throughout the last decade! Buttiglione et al. (2014) : Contrary to a widely held belief, the world has not started to delever
Debt deleveraging (within the EA) By how much has debt decreased /is required to decrease
Debt deleveraging (a current account perspective) Spanish current account adjusted from 10% of GDP in 2008 to 0 in 2013
Collapse in the supply of safe assets