The dollar and real economic activity: an evolving relationship
Burcu Erik, Marco Lombardi, Dubravko Mihaljek and Hyun Song Shin*2020 AEA meetings in San Diego
*The views expressed here are the authors’ and not necessarily those of the Bank for International Settlements
Post-crisis developments in the banking sector
2
3
Total assets and book equity of 17 US banks
Banks: Bank of America Corporation; Bank of New York Mellon Corporation; Capital One Financial Corporation; Citigroup Inc.; CitizensFinancial Group Inc.; Fifth Third Bancorp; JPMorgan Chase & Co.; Morgan Stanley; Goldman Sachs Group Inc.; Northern Trust Corporation;PNC Financial Services Group Inc.; Regions Financial Corporation; State Street Corporation; SunTrust Banks Inc.; Truist Financial Corporation;U.S. Bancorp and Wells Fargo & Company.
0
3
6
9
12
98 01 04 07 10 13 16 19
USD, trn
Total assets Book equity
4
Total assets and book equity of 17 US banks: in log scale
The beta coefficient shows the trend growth rate in natural logarithm. For instance, a 0.15 value corresponds to around 15% growth rate perannum. Banks: Bank of America Corporation; Bank of New York Mellon Corporation; Capital One Financial Corporation; Citigroup Inc.; CitizensFinancial Group Inc.; Fifth Third Bancorp; JPMorgan Chase & Co.; Morgan Stanley; Goldman Sachs Group Inc.; Northern Trust Corporation;PNC Financial Services Group Inc.; Regions Financial Corporation; State Street Corporation; SunTrust Banks Inc.; Truist Financial Corporation;U.S. Bancorp and Wells Fargo & Company.
4
5
6
7
8
9
98 01 04 07 10 13 16 19
βpre = 0.15
βpost = 0.04
βpre = 0.15
βpost = 0.02
Log, USD bn
Total assets Book equity
5
Total assets and book equity of 26 euro area banks
Banks: AIB Group Plc; Banca Monte dei Paschi di Siena SpA; Banco Bilbao Vizcaya Argentaria SA; Banco Bpm SpA; Banco de Sabadell SA;Banco Popular Español SA; Banco Santander SA; Bank of Ireland Group; Bankia SA; BNP Paribas SA; CaixaBank SA; Commerzbank AG; CréditAgricole SA; Crédit Industriel et Commercial SA; Deutsche Bank AG; Dexia SA; Erste Group Bank AG; ING Groep NV and Nordea Bank Abp.;Intesa Sanpaolo SpA; KBC Group NV; Natixis SA; Raiffeisen Bank International AG; Société Générale SA; UniCredit SpA; Unione di BancheItaliane SpA.
0
5
10
15
20
98 01 04 07 10 13 16 19
EUR, trn
Total assets Book equity
6
Total assets and book equity of 26 euro area banks: in log scale
The beta coefficient shows the trend growth rate in natural logarithm. For instance, a 0.15 value corresponds to around 15% growth rate perannum. Banks: AIB Group Plc; Banca Monte dei Paschi di Siena SpA; Banco Bilbao Vizcaya Argentaria SA; Banco Bpm SpA; Banco de SabadellSA; Banco Popular Español SA; Banco Santander SA; Bank of Ireland Group; Bankia SA; BNP Paribas SA; CaixaBank SA; Commerzbank AG;Crédit Agricole SA; Crédit Industriel et Commercial SA; Deutsche Bank AG; Dexia SA; Erste Group Bank AG; ING Groep NV and Nordea BankAbp.; Intesa Sanpaolo SpA; KBC Group NV; Natixis SA; Raiffeisen Bank International AG; Société Générale SA; UniCredit SpA; Unione di BancheItaliane SpA.
4
5
6
7
8
9
98 01 04 07 10 13 16 19
βpre = 0.22
βpost = 0.00
βpre = 0.23
βpost
= –0.03
Log, EUR bn
Total assets Book equity
7
Cumulative dividends since 2007 have eroded bank equity
1066 413 345 180
365
265249
7450
60
70
80
0
400
800
1,200
US EA Other JapanEurope
USD, bn%
Cumultive dividends from 2007 to 2018(DIV, rhs)Retained earnings in 2018(REA, rhs)
Blue / (Blue + yellow) (lhs)
8
Low price-to-book ratios create incentive to pay out equity
0.0
0.5
1.0
1.5
2.0
2007 2010 2013 2016 2019
Bank equity price-to-book ratio
United States Euro area Japan
9
9
Leverage of US securities broker-dealer sector
Leverage1
1 Calculated as total assets divided by equity.Sources: Federal Reserve, Flow of Funds; Bloomberg; BIS calculations.
Leverage and VIX index Leverage and VIX index
Assets/equity Assets/equityLn(VIX)
1
2
3
4
10
20
30
40
99 04 09 14 19
ln(VIX) (lhs) Leverage (rhs)1
10
20
30
40
99 04 09 14 1910
20
30
40
2.5 3.0 3.5 4.0 4.5ln(VIX)
Leve
rage
1
Q1 95–Q1 09 Q2 09–Q2 19
10
10
Leverage of US securities broker-dealer sector and VIX index
1 Calculated as total assets divided by equity.Sources: Federal Reserve, Flow of Funds; Bloomberg; BIS calculations.
1
2
3
4
10
20
30
40
99 04 09 14 19
Ln(VIX) Assets/equity
ln(VIX) (lhs) Leverage (rhs)1
11
11
Leverage no longer dances to the tune of the VIX index
1 Calculated as total assets divided by equity.Sources: Federal Reserve, Flow of Funds; Bloomberg; BIS calculations.
10
20
30
40
2.5 3.0 3.5 4.0 4.5ln(VIX)
Leve
rage
1
Q1 95–Q1 09 Q2 09–Q2 19
12
CIP deviationAvdjiev, Du, Koch and Shin (AER Insights 2019)
1 Average five-year cross currency basis across Australian dollar, British pound, Canadian dollar, Danish krone, euro,Japanese yen, New Zealand dollar, Norwegian krone, Swedish krona and Swiss franc vis-à-vis the US dollar.Sources: Bloomberg; BIS calculations.
–40
–30
–20
–10
0
2001 2003 2005 2007 2009 2011 2013 2015 2017 2019
Basis points
Mean cross-currency basis spread1
13
US dollar broad index and the cross-currency basisAvdjiev, Du, Koch and Shin (AER Insights 2019)
1 Federal Reserve Board trade-weighted nominal dollar index, broad, based on goods trade. Higher values indicate astronger US dollar. 2 Average five-year cross currency basis across Australian dollar, British pound, Canadian dollar,Danish krone, euro, Japanese yen, New Zealand dollar, Norwegian krone, Swedish krona and Swiss franc vis-à-vis theUS dollar.Sources: Federal Reserve Bank of St Louis, FRED; Bloomberg; BIS calculations.
60
80
100
120
–40
–20
0
20
2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
Basis pointsJan 1997 = 100
US dollar index (lhs)1 Mean cross-currency basis spread (rhs)2
14
US dollar credit to EMEs
1 Annual growth of US dollar-denominated credit or loans to non-banks in EMEs. 2 Annual growth of the FederalReserve Board trade-weighted nominal dollar index, major EMEs.Sources: Datastream; Dealogic; Euroclear; FRED; Thomson Reuters; Xtrakter Ltd; national data; BIS locational bankingstatistics; BIS effective exchange rate statistics; BIS calculations.
–20
–10
0
10
20
30
2001 2003 2005 2007 2009 2011 2013 2015 2017 2019
Yoy, %
Total credit1 Loans1 US dollar index2
15
US dollar credit to EMEs
1 Annual growth of US dollar-denominated credit or loans to non-banks in EMEs. 2 Annual growth of the FederalReserve Board trade-weighted nominal dollar index, major EMEs.Sources: Datastream; Dealogic; Euroclear; FRED; Thomson Reuters; Xtrakter Ltd; national data; BIS locational bankingstatistics; BIS effective exchange rate statistics; BIS calculations.
–20
–10
0
10
20
30
2000 2002 2004 2006 2008 2010 2012
Yoy, %
Total credit1 Loans1 US dollar index2
–10
–5
0
5
10
15
2015 2016 2017 2018 2019
Yoy, %
Regression coefficients for bank capital flows
16
Impact of bilateral dollar
appreciation
Impact of broad dollar appreciation
Estimated coefficients from panel regressions, US dollar. Source: BIS calculations.
17
Why broad dollar index?
Consider global lender with diversified portfolio of dollar credits to borrowers around the world
Some borrowers face currency mismatch or otherwise benefit from weaker dollar (eg, oil firm)
Dollar depreciation against whole basket implies: Reduction in credit risk for individual borrowers Reduced tail risk for diversified loan portfolio Reduced Value-at-Risk Increased lending capacity given economic capital
Bruno and Shin (RES 2015)
18
Strong positive relationship between the average basis and• the daily dollar beta (for 3M basis); correlation: 85% (LHP)• the quarterly dollar beta (for 5Y basis); correlation: 97% (RHP)
Evolving relationship between the dollar and PMIs
19
20
Headline, new export orders and employment PMIs1
Headline2 New export orders3 Employment3
1 A value of 50 indicates that the number of firms reporting improvement and deterioration are equal; a value above 50 indicates improvement. Aggregations based on GDP and PPP exchange rates. 2 For manufacturing, eight AEs and 17 EMEs; for services, five AEs and four EMEs. 3 Global.Sources: IMF, World Economic Outlook; Datastream; IHS Markit; BIS calculations.
48
50
52
54
56
2016 2017 2018 2019
Diffusion index
Manufacturing Services
AEs:
EMEs:
49
50
51
52
53
2017 2018 2019
Diffusion index
Services
46
48
50
52
54
2016 2017 2018 2019
Diffusion index
Manufacturing
21
PMIs and financial variables
Purchasing managers closely follow financial markets to assess their financing conditions Correlation between PMIs and high-frequency financial variables Equity prices are good predictors of current-month PMIs (Erik et
al 2019). Principal components from equities, corporate spreads and the USD Focus on β from:
𝑃𝑃𝑃𝑃𝑃𝑃𝑡𝑡 = α + βΔ𝑝𝑝𝑝𝑝𝑡𝑡 + φ𝑃𝑃𝑃𝑃𝑃𝑃𝑡𝑡−1 + ε𝑡𝑡 (1)
22
Nowcasting exercise
Nowcasting PMIs with the first principal component of financial variables
Full sample Pre-GFC Post-GFC 𝑃𝑃𝑃𝑃𝑃𝑃𝑡𝑡−1 0.944* 0.944* 0.943* ∆𝑝𝑝𝑝𝑝𝑡𝑡 0.215* 0.200* 0.077*
R2 0.936 0.910 0.912
RMSE 0.843* 0.906* 0.950*
* denotes results significance at the 1% level. RMSE is computed as the ratio over a plain AR(1) benchmark; significance is determined by Clark and McCracken (2012) test. Full sample: 2/1998 -10/2019 (261 observations); pre-GFC: 2/1998 - 12/2007 (119 observations); post-GFC: 1/2010 - 10/2019 (118 observations).
Sources: Federal Reserve Bank of St Louis, FRED; Bloomberg; Datastream; ICE BofAML indices; IHS Markit; MSCI; authors' calculations.
23
Loadings of the first principal component before and after the GFC
–0.4
–0.2
0.0
0.2
Equity (4) Equity (1) Equity (2) Equity (3) USD VIX Spread (1) Spread (4) Spread (3) Spread (2)
//
Pre-GFC: 2/1998–12/2007 (119 observations)Post-GFC:1/2010–10/2019 (118 observations)
24
Main takeaways from nowcasting exercise
Synthetic index based on PC outperforms the AR(1) benchmark across all sub-periods
Post-crisis PC loading on the dollar higher, at the expense of VIX
Sign on USD is negative, i.e. dollar appreciation acts as a drag on PMIs
25
A small-scale VAR
VAR ordering World equity returns Changes in broad dollar index Global manufacturing PMI (excluding the US) Global trade growth
Equity returns and dollar index changes measured over a window that predates the PMI poll “Global equity price shocks" as news before the PMI release "US dollar shocks" as additional news before PMI release "PMI shock" as change in PMIs not already priced into financial
variables
26
Responses of global trade growth to equity price and US dollar shocks
Pre-GFC: 8/1998 -- 12/2007 (113 observations); post-GFC; 1/2010 -- 6/2019 (114 observations).Equity: MSCI Emerging Markets Index, local currency; US dollar: Federal Reserve Board trade-weighted nominal dollar index, other important trading partners, based on goods trade; Trade: CPBWorld Trade Volume index, seasonally adjusted.
–1.0
–0.5
0.0
0.5
1.0
0 3 6 9 12 15 18 21 24
Pre-GFC Post-GFC
Mean point estimate:
+/- onestandard error:
Months
–1.0
–0.5
0.0
0.5
1.0
0 3 6 9 12 15 18 21 24
Pre-GFC Post-GFC
Mean point estimate:
+/- onestandard error:
Months
Equity prices US dollar index
27
Response of global trade growth to US dollar shocks
US dollar = Federal Reserve Board trade-weighted nominal dollar index, other important tradingpartners, based on goods trade; Trade = CPB World Trade Volume index, seasonally adjusted.
–1.0
–0.5
0.0
0.5
0 3 6 9 12 15 18 21 24
Pre-GFC: 8/1998–12/2007 (113 observations)Post-GFC: 1/2010–6/2019 (114 observations)
Mean point estimate:
+/- one standard error:Months
Perspectives on trade and finance linkage
28
29
Exports volume growth
–12
–6
0
6
12
1984 1989 1994 1999 2004 2009 2014 2019
Yoy, %
AEs EMEs Global
AEs = AU, CA, EU, JP, NZ, NO, SE, CH, DK, GB, US. EMEs = AE, AR, BR, CL, CN, CO, HK, ID, IN, KR, MX, MY, PE, PH, PL, RU,SA, SG, TH, TR and ZA.Sources: IMF, World Economic Outlook; BIS calculations.
30
Trade policy uncertainty1
1 World Trade Uncertainty Index, constructed by counting the number of times uncertainty is mentioned within aproximity to a word related to trade in the Economist Intelligence Unit (EUI) country reports. The index is an equallyweighted average scaled by total number of words in the EUI reports.Sources: Ahir H, N Bloom and D Furceri, "The world uncertainty index"; Brookings analysis of QCEW and EMSI data; BIScalculations.
–10
0
10
20
CA US FR DE CN IT MX JP SA GB TR BR IN RU ZA AR AU ID KR
Index
2011-2017 2018 2019
31
Exports value growth1
–12
–8
–4
0
4
ID AU IN MX AR FR TR CA US EU IT JP GB DE CN ZA RU BR SA KR
Yoy, %
2011-2018 2019
1 Simple average of growth rates for periods shown.Sources: IMF, World Economic Outlook; BIS calculations.
32
1 Both exports and GDP are measured in constant prices.Sources: IMF, World Economic Outlook; World Trade Organization; Datastream; national data; BIS calculations.
Ratio of world goods exports to world GDP
95
100
105
110
115
2001 2003 2005 2007 2009 2011 2013 2015 2017 2019
Q1 2000 = 100
World goods exports / GDP1
33
1 Federal Reserve Board trade-weighted nominal dollar index, broad group of major trading partners of the US (“broad”), based only on trade in goods. An increase indicates appreciation of the US dollar. 2 Both exports and GDP are measured in constant prices.Sources: Federal Reserve Bank of St Louis, FRED; IMF, World Economic Outlook; World Trade Organization; Datastream; national data; BIS calculations.
Global goods trade and the dollar
80
87
94
101
108
95
100
105
110
115
2001 2003 2005 2007 2009 2011 2013 2015 2017 2019
Q1 2000 = 100 Q1 2000 = 100
US dollar index (lhs)1 World goods exports / GDP (rhs)2
34
Findings from empirical study using micro data
Micro study of Mexican exporters Loans matched with borrowing firm and lending bank 4.6 million shipments by destination and 8 digit product category
Findings Exporters who borrow from banks more dependent on dollar
wholesale funding exhibit greater contraction of export shipments when the broad dollar appreciates
Evidence for both working capital channel and dollar invoicing channel (Gopinath and Stein (2019), Gopinath et al. (2019))
Bruno and Shin (2019) “Dollar exchange rate as a credit supply factor: evidence from firm-level exports” BIS Working Paperhttps://www.bis.org/publ/work819.htm
35
Broad dollar index
Broad dollar index has attributes of an indicator of dollar funding conditions
Indicator of shadow price of bank balance sheet capacity
Possible role as linchpin that ties together financial and real variables
