Discussion of “ Capital Flows and the Adjustment to Common Shocks in a Two-Country Business Cycle Model ” Ivan Jaccard & Frank Smets Robert Kollmann, ULB and CEPR Bank of France, December 13, 2013 1
● IMPORTANT questions. ● Paper documents interesting stylized facts about European biz-cycle: core vs. periphery ● Tries to explain different biz-cycle dynamics across EA countries as outcome of structural differences ● Argues that less efficient financial intermediation in Euro periphery is key ● Innovative, thought-provoking ● Careful and well-written paper 2
Stylized facts about Euro Area Biz cycles, 1999-2012, CORE vs PERIPHERY (nco: net capital outflow = current account, CA) ● Relative volatility of C is higher in periphery ● CORE has CA surplus; Periphery runs CA deficit ● CORE CA PRO-CYLICAL; Periphery CA is Counter-cyclical ● Real short term interest rate (rr) lower in periphery ● Lending rate spread L r r higher in periphery D 3
4
Ivan & Frank ’ s model: ● Two-country RBC model with banks ● Each country has a bank that collects deposits locally and makes loans locally ● Each bank needs (some) capital provided by foreign bank ● Technology of bank: are ‘ produced ’ using labor (monitoring of borrowers) y N Loans , L t , B t , d , foreign bank capital b deposits from local households t t 5
► Question: why are date t deposits an input into production of gross change of stock of loans? More plausible that deposits and bank capital at t are input for date t STOCK of loans : Loans(t) = Deposits(t) + Bank equity (t) ● Final-good producing sector: 1 uses loans as PHYSICAL input y A y N t t L t , F t , Short-cut for generating effect of banking on real activity ● Just one type of shock: TFP in Home and Foreign final good production (TFP identical) ● Cross-country differences: risk aversion, capital adjustment cost, monitoring intensity in banking 6
● Paper estimates country specific parameters using SMM: Periphery has greater capital adjustment costs, greater monitoring costs, but lower habit persistence than CORE. ● Real interest rate difference Core vs. periphery used to pin down risk aversion in Core vs. periph. Question: how do you solve model? up to linear approximation, mean interest rate is pinned down my rate of time preference only. Paper should explain numerical solution method. What determines steady state NX in this model? Paper should explain determinants of SS. Is SS unique? 7
● Parameter asymmetries across countries explain why a common TFP shock affects the trade balance. A positive (common) TFP shock rise in Core Net Exports, fall in periphery NX Authors argue that this mainly reflects greater financial monitoring intensity in periphery. 8
Counterintuitive: would expect that greater monitoring intensity in periphery limits expansion of periphery investment after a (common) TFP shock. Positive TFP shock raises wage rate; greater labor intensity of periphery banking means that lending rate spread rises more in periphery. This dampens expansion of periphery investment. Hence periphery absorption will rise less, on impact Periph. NX ; Core NX 9
Future versions of paper should provide more intuition for this key result Drawback of having only 1 shock: output, C and I will be (almost) perfectly correlated across countries. The authors do not report actual and predicted cross-country correlations. In reality: Core had lower growth than periphery during the boom. Core has higher growth in aftermath of crisis. The model here cannot explain these asymmetries. 10
MAIN COMMENT 1999-2012 : short sample with major structural breaks and very big & unusual shocks: ● Convergence to new steady state after major regime change: creation of Euro; increase in international financial integration (elimination of exchange rate uncertainty; ‘ Markets in Financial Instruments Directive ’ , MiFID, of 2004) ● Global financial crisis ● Boom-bust cycle in periphery 11
1999-2012: initial conditions matter; sample moments are NOT representative of stationary distribution of simple RBC models ● Relative volatility of C is high in periphery. Probably not structural. Due to collapse of durable good consumption in periphery (sudden stop)? ● CORE runs trade balance surplus; Periphery runs TB deficit Most likely ‘ conjunctural ’ (adjustment to Euro), not structural as in model ● CORE NX PRO-CYLICAL; Periphery NX is Counter-cyclical Not structural (see below) ● Real short term interest rate (rr) lower in periphery. Due to higher inflation in periphery (boom triggered i.a. by foreign capital inflow & housing bubble after launch of Euro) [not in model] ● Lending rate spread r r higher in periphery. Probably reflects L D banking crisis in periphery during 2 nd part of sample [not in model] 12
Counter-cyclicality of net exports (NX) is one of most robust stylized facts E.g. Backus, Kehoe & Kydland (AER, 1994) 13
Facts about European External Accounts ● Kollmann, Ratto, Roeger, in ’ t Veld & Vogel (2013), " What drives the German current account? And how does it affect other Euro Area member states? " Prepared for April 2014 Economic Policy Panel ● in ’ t Veld, Kollmann, Pataracchia, Ratto & Roeger (2013), " International Capital Flows and the Boom-Bust Cycle in Spain " 14
Corr German CA & NX with German GDP (relative GDP) Corr between German CA & Corr between German NX & DE DE REA DE ROW DE DE REA DE ROW Y Y / Y Y / Y Y Y / Y Y / Y (1) (2) (3) (4) (5) (6) (a) 1991-2012 dtrendL -0.13 -0.28 -0.59 0.11 -0.54 -0.10 fdiff -0.00 -0.21 -0.23 0.25 -0.02 0.10 HP -0.07 -0.55 -0.56 0.08 -0.52 -0.24 (b) 1960-1990 dtrendL -0.18 0.03 -0.95 -0.41 -0.32 -0.88 fdiff -0.36 -0.36 -0.51 -0.38 -0.22 -0.06 HP -0.18 0.14 -0.40 -0.28 -0.08 -0.18 (c) 1960-2012 dtrendL -0.30 -0.29 -0.71 -0.47 -0.47 -0.69 fdiff -0.29 -0.50 -0.50 -0.23 -0.43 -0.33 HP -0.35 -0.48 -0.73 -0.40 -0.56 -0.66 15
Spectacular increase of German CA surplus after launch of Euro (1999) ● slightly negative 1991 -2000 (-1% GDP) ● strong rise 2002 -2007 (+7.5% of GDP) ● stable 2008 -12 (5%-6% of GDP) ● in same league as CAs of Japan & China 16
CA = trade balance + net transfers & income CA = S – I 17
S/Y: falls until 2003, then I/Y: trend 18
19
Fall in I/Y: common to different sectors 20
Net Exports: NX=Y-C-G-I NX/Y=(Y-C-G)/Y – I/Y 21
CA surplus of Germany, REA and ROW German CA surplus & REA CA deficit Bilateral German CA with REA & ROW German CA surplus with REA & REA CA deficit 22
23
24
25
Kollmann et al. (2013) estimate rich New Keynesian three-country model using data for Germany, Rest of Euro Area & ROW 26
RESULTS: No mono-causal explanation of DE CA German CA surplus driven by succession of shocks: 1) Increase in financial integration among EMU members convergence of interest rates Rest of Euro Area interest rates DE interest rate Investment in DE ; GDP in DE Invest & GDP in Rest of EA [Less important than other drivers] 27
2) Growth in emerging economies: DE exports 3) Labor market reforms & social security reform in DE: DE labor supply wage restraint in DE DE competitiveness DE GDP ; REA GDP (weak) 4) Depressed demand: high household saving rate (related to population ageing + pension reform) low investment rate 28
● Shocks that drove the rise in German CA Lowered REA CA balance, but had WEAK effect on GDP in Rest of Euro Area 29
Historical Decomposition: German Net Exports, as % of GDP Interest convergence Data Labor mkt reform Technology External Saving Shock Demand 30
German GDP, year-on-year growth rate Labor market reform Technology 31
German inflation (year-on-year growth of GDP deflator) 32
REA net exports divided by nominal GDP ROW demand 33
Summary: Very useful exercise: How do structural differences of countries affect cyclical behavior of int ’ l capital flows Future versions: should explore model with richer structure of shocks and initial conditions (transition to new regime after launch of Euro) Look forward to reading future papers by Ivan and Frank on these important issues 34
THANK YOU ! 35
Recommend
More recommend
