(VII) INTERNATIONAL INTEGRATION OF FINANCIAL MARKETS

Transcription

1 (VII) INTERNATIONAL INTEGRATION OF FINANCIAL MARKETS LECTURES Question 1: What are the arguments in favor of open financial markets? Question 2: Does it really work this way? Question 3: How integrated are financial markets, and what are the remaining barriers?

2 Advantages of financial opening For a successfully-developing country, with high return to domestic capital, investment can be financed more cheaply by borrowing abroad than out of domestic saving alone. Symmetrically, investors in rich countries can earn a higher return on their saving by investing in emerging markets than they could domestically. Households can smooth consumption over time. In the presence of uncertainty, investors can diversify away some risks.

3 future wine Classic gains from trade Under free trade, Portugal responds to new relative prices by shifting into wine, where it has a comparative advantage. In autarky, Portugal can only consume what it produces. (Price mechanism puts it on full-employment PPF & at the point maximizing consumers utility.) Textiles are cheaper on world markets. Portuguese consumption in textiles rises, which it imports, thereby reaching a higher indifference curve. today textiles Next, we do the gains from trade again, substituting period 0 & period 1, in place of wine & textiles.

4 Intertemporal optimization We will maximize the intertemporal utility function: u[c 0 ] + β u[c 1 ] where C 0 consumption today; C 1 consumption tomorrow; u'(c) > 0; u''(c) < 0; β subjective discount factor, reflecting patience. 0<β<1. Total resources available = Y r Y 1. where Y 0 income today; Y 1 income tomorrow; r real interest rate. Total spending discounted to today = C r C 1. Budget constraint: C 1 = (1+r)(Y 0 -C 0 ) + Y 1. Intertemporal utility subject to budget constraint: u[c 0 ] + β u[(1+r)(y 0 -C 0 ) + Y 1 ] To maximize, differentiate with respect to C 0 : Euler equation: u'[c 0 ] + β u'[c 1 ](-(1+r)) = 0. u'[c 0 ]/u'[c 1 ] = β (1+r)

5 A simple functional form for intertemporal utility Let s try the case of log utility: log[c 0 ] + β log[c 1 ] (a special case of iso-elastic utility functions) Then Euler equation u'[c 0 ]/u'[c 1 ] = β (1+r) becomes [1/C 0 ]/[1/C 1 ] = β (1+r). => C 1 C 0 = β (1+r). Result: Agents choose higher consumption tomorrow than today if r is high and/or they are patient. Aggregating up, gives a theory to determine the interest rate: 1+r = C 1 C 0 / β => r > 0 if opportunities grow over time and people are impatient.

6 Welfare gains from open capital markets: 1. Even without intertemporal reallocation of output, Y 0 & Y 1, consumers are better off (borrowing from abroad to smooth consumption). 2. In addition, firms can borrow abroad to finance investment. The intertemporal optimization theory of the current account WTP, 2007 Prof.J.Frankel

7 The intertemporal-optimization theory of the current account, and welfare gains from international borrowing 1. Financial opening with fixed output => domestic residents borrow from abroad, so that they can consume more in Period 0. future Y 1 High interest rate encourages agents to postpone consumption. Assume interest rates in the outside world are closer to 0 than they were at home (the slope of the line Y 0 is closer to -1.0). => C 0 Welfare is higher at point B. today Source: Caves, Frankel & Jones (2007) Chapter 21.5, World Trade & Payments, 10 th ed. Prof.J.Frankel

8 The intertemporal optimization theory of the current account, and welfare gains from international borrowing, continued 2. Financial opening with elastic output Assume interest rates in the outside world are closer to 0 than they were at home. future Shift production from Period 0 to 1, and yet consume more in Period 0, thanks to foreign capital flows. Welfare is higher at point C. today Source: Caves, Frankel & Jones (2007) Chapter 21.5, World Trade &Payments.

9 Does this theory ever work in practice? Norway discovered North Sea oil in 1970s. It temporarily ran a large CA deficit, to finance investment (while the oil fields were being developed) & to finance consumption (as was rational, since Norwegians knew they would be richer in the future). } } Subsequently, Norway ran big CA surpluses. ITF220 Prof.J.Frankel

10 Effect when countries open their stock markets to foreign investors, on cost of capital. Peter Henry (2007) Capital Account Liberalization: Theory, Evidence, and Speculation, JEL, 45(4): Liberalization occurs in Year 0. Cost of capital falls, on average.

11 Effect when countries open their stock markets to foreign investors, on investment. Peter Henry (2007) Capital Account Liberalization: Theory, Evidence, and Speculation, JEL, 45(4): Liberalization occurs in Year 0. Investment rises, on average.

12 Indications that financial markets do not always work as advertised 1) The Lucas Paradox 2) Pro-cyclical capital flows 3) Crises

13 Indications that financial markets do not always work as advertised 1) The Lucas paradox: Capital flows do not systematically go from rich countries (high K/L) to poor (low K/L). Robert Lucas (1990), Why Doesn t Capital Flow from Rich to Poor Countries? AER. Capital flows uphill. Possible explanation: In many developing countries investors cannot reap the potential returns to capital due to inferior institutions, especially inadequate protection of property rights. -- Alfaro, Kalemli-Ozcan & Volosovych (2008).

14 Indications that financial markets do not always work as advertised 2) Pro-cyclicality: Capital flows tend to be pro-cyclical, not counter-cyclical. E.g., Kaminsky, Reinhart & Végh (2005) When it rains, it pours. Possible explanations: In developing countries, (i) given imperfect creditworthiness, investors require collateral, e.g., tangible foreign exchange earnings. The value of the collateral is higher in booms than busts. (ii) Fluctuations that appear cyclical, in truth may signal changes in long-run growth prospects. -- Aguiar & Gopinath (2007).

15 Indications that financial markets do not always work as advertised 3) Crises Debt crises, currency crises, banking crises The 1982 international debt crisis; crisis in the European Exchange Rate Mechanism; EM currency crashes of the late 1990s: Mexico; 1997 E.Asia, esp. Thailand, Korea & Indonesia; 1998 Russia, 2000 Turkey, 2001 Argentina, 2002 Uruguay GFC (U.S. & U.K.: North Atlantic Financial Crisis!) Iceland, Hungary, Latvia, Ukraine, Pakistan ; The euro crisis (Greece, Ireland, Portugal, Spain, Cyprus ).

16 Indications that financial markets do not always work as advertised, cont. Do investors punish countries when and only when governments follow bad policies? Large inflows often give way suddenly to large outflows, with little news appearing in between to explain the change in sentiment. Contagion sometimes spreads to countries that are unrelated, or where fundamentals appear stronger. Recessions have been so big, it seems hard to argue that the system works well.

17 Empirical studies of financial openness and economic performance, reviewed by Kose, Prasad, Rogoff & Wei (2009), often find little systematic relationship, in either direction.

18 Some studies find that financial openness is helpful only if countries have already attained an adequate level of: income -- Biscarri, Edwards, & Perez de Grarcia (2003); Klein & Olivei (1999); Edwards (2001); Martin & Rey (2002); Ranciere, Tornell & Westermann (2008); financial depth, institutional quality & other reforms -- Kaminsky & Schmukler (2003); Chinn & Ito (2002); Klein (2003); Obstfeld (2009); Kose, Prasad & Taylor (2009); Wei & Wu (2002); Prasad, Rajan & Subramanian (2007). Or macroeconomic discipline. -- Arteta, Eichengreen & Wyplosz (2001). => Conventional wisdom regarding sequencing: it is better to liberalize financial markets only after other reforms have been put in place. -- McKinnon (1993), Edwards (1984, 2008), and Kaminsky & Schmukler (2003).

19 Measuring International Financial Integration I. Direct measures of barriers, e.g., IMF s count of freedom from KA restrictions. II. Price tests III. Quantity tests e.g., intl. assets+liabilities/gdp Source: Kose, Prasad, Rogoff & Wei (2009)

20 Menzie Chinn & Hiro Ito, "A New Measure of Financial Openness" (Journal of Comparative Policy Analysis, 2008), updated July I. Direct Measure of Financial Barriers: Chinn-Ito tally of capital controls, from IMF data Figure 1: Development of KAOPEN for Different Income Groups Development of KAOPEN for different income groups Rapid financial liberalization in 1990s year Industrial Countries Emerging Markets Less Developed

21 Chinn-Ito Measure of Financial Openness The calculations are based on 4 categories in the IMF s Annual Report on Exchange Arrangements & Exchange Restrictions: multiple exchange rates, current account restrictions, capital account restrictions, and required surrender of export proceeds.

22 Measuring International Financial Integration, cont. II. Price tests 1.Uniform price of an asset across markets E.g., gap between China s A shares and off-shore shares. 2. Interest rate parity (IRP) i) Covered interest parity (CIP) ii) Uncovered interest parity (UIP) iii) Real interest parity (RIP)

23 1. Price of the same asset across borders Chinese firms stock prices remain higher onshore than offshore. Premium of A shares (held in Shanghai), over H shares (held in Hong Kong) } June 2014: After net inflows turn to outflows, reserves start to fall. Nov. 17, 2014: Shanghai-HK Stock Connect goes into effect. Nov. 21: PBoC starts cutting interest rates. Bubble follows. Summer 2015: Shanghai stock market & RMB begin to fall; China s authorities pressure domestic institutions to buy stocks. Source: FT, Feb.29, 2016 Prof.J.Frankel

24 I. Currency factors 2. Interest Rate Parity: Why does i not equal i*? Expected currency depreciation Exchange risk premium The currency premium can be measured as the forward discount, or swap rate, or differential between domestic & local $-linked bonds. II. Country factors

25 Decomposition of the nominal interest differential i i* country premium + currency premium e.g., ( i i* - fd ) + fd fd (fd - Δs e ) + (Δs e ) exchange risk premium + expected nominal depreciation The country premium could be measured by the sovereign spread, Credit Default Swap, or covered interest differential (i-i*-fd). The currency premium could be measured by the forward discount (fd), currency swap rate, or local spread of $-linked vs. domestic-currency bonds.

26 WHY DOES i NOT EQUAL i*? II. Country factors, continued Default risk reflected in sovereign spreads or Credit Default Swaps Capital controls reflected in covered interest differentials Taxes on cross-border investments Transaction costs Imperfect information Risk of future capital controls

27 Sovereign spreads Brazilian interest rate decomposed country premium + currency premium + LIBOR } } Total spread (Brazil rate minus LIBOR) = Currency premium (forward premium) + Country premium (spread)

28 Sovereign spreads Mexican spread decomposed: currency premium + country premium Total spread for Mexican sovereign bonds over US Treasury bill interest rate Currency premium pesos/$ swap rate Country premium total spread adjusted for currency premium } Total spread over US T bill rate Country premium Currency swap rate Wenxin Du & Jesse Schreger, Sovereign Risk, Currency Risk & Corporate Balance Sheets, Oct. 2014

29 EMBI+ Sovereign spreads, Spreads were low for Emerging Market bonds in 2006, and even lower for South Africa EMBI+ 150 RSA EMBI+ } Jun Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul- 06 Global investors were under-pricing risk -- as also reflected in US corporate spreads, options prices, etc. All of them shot back up in 2008.

30 Sovereign spreads for 5 euro countries shot up in the 1 st half of 2010

31 Snapshot of spot & forward exchange rates on Aug. 14, 2014 Spot rate Transaction cost Forward rates Selling at a forward discount vs. the $ * * Selling at a forward premium * Value of and are shown in terms of $. All other exchange rates are local per $.

32 Spot and forward exchange rates on Aug. 14, 2014, continued Spot rate Transaction cost Forward rates Selling at a forward discount vs. the $ $ pegs

33 => ( 1 + i Turkey ) COVERED INTEREST PARITY = (1/S) ( 1 + i US ) where S is the spot rate in TL/$ and F is the forward rate. (1 + i Turkey ) = (1 + fd) (1 + i US ). = (1 + fd + i US + fd i US ). Forward discount fd (F-S)/S => 1 + fd F/S Because (fd i US ) is small, i Turkey fd + i US. => If the Turkish nominal interest rate exceeds the U.S. rate, then the lira sells at a discount in the forward exchange market. F

34 Liberalization in a country that had controls on capital inflows. Domestic & offshore interest rates, Germany, } From: Marston (1989)

35 Liberalization in a country that had controls on outflows. France, : Domestic and Offshore Interest Rates { From: Mussa and Goldstein,(1993). France kept its controls on capital outflows until the late 1980s. They produced an offshore-onshore differential, which shot up whenever there was speculation of a franc devaluation. Again, the differential disappeared after controls were removed. ITF220 Prof.J.Frankel

36 In late 2008 Covered Interest Parity surprisingly failed, in the Global Financial Crisis rush to the $ as safe haven. Covered interest differentials, using Overnight Index Swap interest rates, Significant determinants are apparently counterparty risk & liquidity, proxied by financial stock CDS, VIX, implied fx volatility, OIS bid-ask spreads & Fed swap lines. Inês Isabel Sequeira de Freitas Serra, Covered Interest Parity, NOVA School of Business & Economics, Lisbon, Jan

37 THREE INTEREST RATE PARITY CONDITIONS Investors decide whether to hold: Arbitrage => parity condition. Does it hold in practice? CIP Covered interest parity $ deposits in New York vs. covered deposits in London i $ NY - i L = fd. Yes, if default risk & capital controls are low & liquidity high. UIP Uncovered interest parity $ deposits in NY vs. deposits in London uncovered. i $ NY - i L = Δs e If risk is unimportant. Hard to tell in practice. RIP Real interest parity Arbitrage is not directly relevant i $ NY - i L = π US e - π UK e No, not in short run.

38 Summary of Interest Rate Parity conditions to be used in L24-26: Exchange Rate Models } Covered interest parity i i* = fd + No risk premium fd = Δs e => Uncovered interest parity i i* = Δs e, + Ex ante Relative Purchasing Power Parity Δs e = π e π* e => Real interest parity i i* = π e π* e. }

39 III QUANTITY TESTS: some show rising integration IMF

40 Quantity tests point to surprisingly low international integration 1. Home bias in portfolios: Do citizens of each country hold a basket of assets that is optimally diversified internationally? 2. Consumption risk-sharing: Are countries consumption levels correlated with each other more than country incomes? 3. Feldstein-Horioka test: Do countries Investment rates vary independently of their National Saving rates? No No No

41 Feldstein-Horioka test of capital mobility Regression: (I/GDP) = α + β (NS/GDP) + v. Feldstein (1980) argued that if capital were perfectly mobile, we would find β = 0: countries with good investment opportunities could borrow abroad to finance them. Instead, β was much closer to 1: Countries are apparently savings-constrained.

42 The Feldstein-Horioka, still as high as 0.7 in the 1980s, declined in the 90s and until Kristin Forbes, Financial deglobalization?: Capital flows, banks, and the Beatles, Bank of England, 18 Nov., 2014

43 Appendices: Country risk Appendix 1: Inter-shuffling of credit-worthiness between advanced & developing countries Recent credit rating rankings The end of original sin? Appendix 2: EM Sovereign Spreads More examples Risk on risk off

44 Appendix 1: The blurring of lines between debt of advanced countries and developing countries 1) Since the crisis of the euro periphery began in Greece in 2010, we have become aware that advanced countries also have sovereign default risk. 2) After 2000, Emerging Market Countries increasingly became able to borrow in their own currencies, so their debt carries currency risk (not just default risk).

45 1) Country creditworthiness became inter-shuffled Advanced countries EM & Developing countries AAA Germany, UK Singapore, Hong Kong AA+ US, France AA Belgium Chile AA- Japan China A+ Korea A Malaysia, South Africa A- Brazil, Thailand, Botswana BBB+ Ireland, Italy, Spain BBB- Iceland Colombia, India BB+ Indonesia, Philippines BB Portugal Costa Rica, Jordan B Burkina Faso SD Greece S&P ratings, 2012

46 Spreads for Italy, Greece, & other Mediterranean members of were near zero, from 2001 until 2008, and then shot up in Market Nighshift Nov. 16,

47 2) The end of Original Sin? After 2000, Emerging Markets successfully issued more debt in their own local currencies (LC), instead of $-denominated (FC). Fig. 2 from Jesse Schreger & Wenxin Du Local Currency Sovereign Risk, HU, March 2013

48 Turkey is able to borrow in local currency (lira), but has to pay a high currency premium to do so. { Total premium on Turkey s lira debt over US treasuries Pure default risk premium on lira debt { Fig. 5 from Schreger & Du, Local Currency Sovereign Risk, HU, March 2013

49 Appendix 2: EM sovereign spreads Sovereign spreads Spreads shot up in 1990s crises EMBI,

50 10/15/1996 2/15/1997 6/15/ /15/1997 2/15/1998 6/15/ /15/1998 2/15/1999 6/15/ /15/1999 2/15/2000 6/15/ /15/2000 2/15/2001 6/15/ /15/2001 2/15/2002 6/15/ /15/2002 2/15/2003 6/15/ /15/2003 2/15/2004 6/15/ /15/2004 2/15/2005 6/15/ /15/2005 2/15/2006 6/15/2006 Downtrend in SA country risk premium, to below 100 basis points by 2006, in tandem with upgrades by rating agencies Sovereign spreads Sovereign spreads on South African Dollar Debt S&P Upgrade (BB+ to BBB-) Source: SA Treasury S&P Upgrade (BBB- to BBB) Moody's upgrade (Baa3 to Baa2) S&P Upgrade (BBB to BBB+) Moody's upgrade (Baa2 to Baa1) Global 06 Global 09 Global 12 Global 14 Global

51 EM sovereign spreads Sovereign spreads Spreads fell to low levels by WesternAsset.com

52 EM sovereign spreads Sovereign spreads Spreads rose again in Sept. 2008, especially on $- denominated debt Bpblogspot.com particularly in Eastern Europe. World Bank

53 What determines spreads? Sovereign spreads EMBI is correlated with risk perceptions risk on risk off Data sources: Bloomberg and Federal Reserve Laura Jaramillo & Catalina Michelle Tejada, IMF Working Paper, March 2011