Towards a New Monetary Theory of the Exchange Rate

Transcription

1 Towards a New Monetary Theory of the Exchange Rate Ambrogio Cesa-Bianchi, BoE Andrej Sokol, BoE Michael Kumhof, BoE Greg Thwaites, BoE

2 November 1, 17 The views expressed herein are those of the authors and should not be attributed to the Bank of England.

3 1 Introduction The Monetary Approach to Exchange Rates: Popular in the 7s/8s: Boughton (1988), Frenkel and Mussa (1985). Key Ingredients: Money demand: Function of output and risk-free interest rate. Money supply: High-powered money only. PPP (in the long run). UIP between risk-free interest rates. Rational expectations. Exogenous risk-free interest rates. Empirically not successful. Reasons: See Key Ingredients. Key theoretical weaknesses from our point of view: Treats risk-free interest rates as exogenous: No policy rules. Treats bank liabilities as securities: 97% of all broad money.

4 The Portfolio Balance Approach to Exchange Rates: Popular around the same time: Branson and Henderson (1985). Similar ingredients. Difference: Relative securities supplies enter UIP. Empirically only slightly more successful. The UIP Approach to Exchange Rates: Popular now. Policy rules for risk-free interest rates. Asset markets only feature risk-free securities. Money demand and money supply are completely ignored. Bank deposit money is typically also completely ignored. Empirically not successful either.

5 A New Monetary Approach to Exchange Rates - Key Ingredients: (differences to old Monetary Approach shown in bold) Money demand: Function of activity and deposit interest rates. Money supply: Bank deposit money only. PPP (in the long run). Modified UIP between deposit interest rates: Exchange rate = relative price of two monies. The monies are produced endogenously by banks. They are imperfect substitutes: Monetary wedge in UIP. Rational expectations. Risk-free interest rates determined by policy rules.

6 The key difference is therefore a different treatment of assets. Monetary Approach: Money = exog. medium of exchange,i m =. UIP Approach: No money. Securities = pure store of value. New Monetary Approach: Money = Medium of exchange and store of value,i d >. Demand for money: Increasing in the deposit rate. Supply of money: Endogenous, created by banks. Similarities with Portfolio Balance Approach: Asset stocks enter the UIP condition. But because of monetary features, not because of risk. Most interest-bearing assets are treated as money, not as securities.

7 The Model - General Considerations.1 Overview Two-countries, Home and Foreign. Households and banks. No government, no investment. Output produced by land (=collateral) and labor. Sticky prices and wages. PCP in baseline, LCP possible. Monetary policy: IFB rule. No fiscal sector. Home and Foreign goods: CES goods demand. Home and Foreign currencies: CES liquidity function. Banks alone create the currencies: Banks create both domestic and foreign currencies. Domestic currencies supported by lender of last resort. Foreign currencies supported by interbank liquidity. Banks create currencies for both households and foreign banks. Cross-border lending only at the interbank level.

8 Home country Banks Interbank deposits (H, F) Foreign country Banks* Deposits (H, F) Lending (H, F) Interbank lending (H, F) Deposits (H, F) Lending (H, F) Households Trade in goods Households*

9 . Endogenous Money in DSGE Models Sidrauski-Brock monetary models of the 198s/199s: 1. Representative household with a demand for money.. Government exogenously supplies all money. The main problem is, not 1. Therefore, in our model: Bank deposits (97% of all money) enter into TA cost technology. Government money is omitted entirely. How are bank deposits created? Never through intermediation of existing funds. Always through new loans (and asset purchases). Jakab and Kumhof (15), Goodfriend and McCallum (7).

10 .3 Banks are not Intermediaries of Loanable Funds The Loanable Funds Model - Postulated Credit Process Intermediation = Physical Trading of Real Resources Banks collect a deposit of real resources (goods or capital) from a saver. Banks lend those existing goods to another agent, the borrower. Deposits in this model are an input. Money in this model is held as a store of value. Rapid changes in credit: Switching of savings, dis-/re-intermediation. The Financing Model Actual Credit Process Financing = Digital Creation of Monetary Purchasing Power Banks make a loan of money to agent X. Banks credit new money to the deposit account of the same agent X. Deposits in this model are an output. Money in this model is held as a medium of exchange. Rapid changes in credit: Changes in gross balance sheet positions. This financing process is a fact, not a theory that needs to be proved.

11 5.6 Models 1: Budget Constraints Budget Constraints in Model 1: Saver Household: deposits s t = income s t spendingt s Borrower Household: capitalt b q t loans b t = incomeb t spendingb t Budget Constraint in Model : Representative Household: capital r t q t+ deposits r t loansr t = incomer t spendingr t

12 3 The Model - Details 3.1 Banks Loans: Bernanke, Gertler and Gilchrist (1999) Costly state verification. Difference: Pre-committed lending rates. Collateral for loans in either currency: Land. 4 loan types: (domestic/foreign ccy) x (non-banks/banks). Deposits: Schmitt-Grohé and Uribe (4) Transactions cost technology. Difference: Money = H deposits + F deposits. 4 loan types: (domestic/foreign ccy) x (non-banks/banks).

13 Monetary Distortion = Liquidity Taxes: τ liq t = 1 +s t +s t v t Equivalent to consumption taxes. Banks effect on the real economy: Through these taxes. Not through intermediation of loanable funds. The Liquidity-Generating Function: f t = (1 γ) 1 ǫ 1 ǫ Deposits H ǫ t +γ 1 ǫ ǫ ǫ 1 ǫ 1 Deposits F ǫ t

14 3. Bank Regulation 3..1 Capital Adequacy Regulation (CAR) Minimum capital adequacy ratio (MCAR) = 8%. Penalty for violations of this ratio. Continuum of banks: Ex-ante low probability of violations. Banks self-insure by keeping more than 8% capital. CAR is therefore never exactly binding. 3.. Currency Mismatch Regulation (CMMR) Bank asset and liability currency exposures to non-banks must match. This corresponds to regulation in many jurisdictions. It pins down balance sheet positions.

15 3.3 Bank Frictions Cross-Border Interbank Liquidity Banks can and do create deposits in foreign currency. But they cannot access a LoLR in foreign currency. Instead they procure foreign currency liquidity from foreign banks. This is costly and affects their pricing of foreign loans. Foreign banks create this liquidity in the same way as for non-banks Gross Foreign Assets Position Cost Cost if domestic currency interbank lending and deposits are mismatched. Limits the extent of currency mismatch on interbank positions. Cannot eliminate mismatch: NFA adjustments through interbank market. In Schmitt-Grohe and Uribe (3) such a cost is applied to net positions. In our model, such a cost must be applied to gross positions.

16 3.4 Households Preferences: maxe β t t= Consumption aggregator: c t = 1 ν x log (c t (j) νc t 1 ) ψ h t (j) 1+1 η η (b c θc 1 H )1/θ c c θc H,t + (1 b c θc 1 H )1/θ c c θc F,t Money demand - transactions cost: s h t =A h čt ď liq or s h t =A h ˇq t land t ď liq t Liquidity aggregator: ď liq t = b d H 1/θd ďh,t θ d 1 θ d + 1 b d H θc θc 1 1/θd θ d 1 et ď F,t θ d θ d 1 θ d

17 Budget constraint: Take out loans (each ccy) to obtain deposits (each ccy). Use land as collateral. Consume, supply labour. Optimality: Monetary wedges for consumption (or land). Monetary wedges for deposits: Wedges enter the UIP condition!

18 3.5 Balance of Payments Current account equation: ľ b H,t +e tďb F,t ďb H,t e tľb F,t = 1 x r b lh,t ľ b H,t 1 +rb df,t e t 1ďb F,t 1 rb dh,tďb H,t 1 rb lf,t e t 1ľb F,t 1 +e t p 1 n H,tč H,t n p F,tč F,t All cross-border positions are gross interbank positions. This is a reasonable assumption in light of the data: Next page. NFA are therefore a netted position between 4 interbank balances. Nobody makes decisions on net positions, only on gross positions. The net position is simply an accounting residual.

19 UK balance sheet positions in 15 (Bowers (16)): The vast majority of cross-border positions involves financial institutions. Households and firms are much less important.

20 3.6 Calibration Mostly US data, except external currency composition. Rest of the world assumed to be of equal size and symmetric (for now). Steady State Real Interest Rates in the Model 5.% = Retail Lending Rates (both ccies) 3% = Policy Rate 4.% = Wholesale Lending Rate (for. ccy) 3.6% = Wholesale Lending Rate (dom. ccy) 3.3% = Interbank Lending Rates (both ccies).9% = Interbank Deposit Rates (both ccies) 1.9% = Retail Deposit Rates (both ccies) %

21 Bankruptcy and MCAR violation rates calibrated from literature. Interbank deposits / GDP: 5%. Consumption: Foreign goods consumption / total goods consumption: 14%. Elasticity of substitution: θ c = 1.5. Deposits: Foreign currency deposits / total deposits: 6%. Elasticity of substitution: θ d = sensitivity analysis. Total lending / GDP: 1%.

22 4 Simulation Results

23 Contraction in Bank Money Supply 1 Comparison: Goods Market versus Asset Market Money Demand (magnitude: 1% of GDP) GDP Base Alt1 Alt Home Inflation Real Policy Rate Foreign Inflation Depreciation. Real Exchange Rate Consumption Exports/GDP Imports/GDP Terms of Goods Trade 1 Share of Foreign Goods in Consumption Current Account/GDP Terms of Money Trade (RR_D_H/RR_D_F) Share of Foreign Ccy Deposits in Total Deposits Price of Land Retail Lending Margin in Home Currency.6.6 Home Currency Loans/GDP Home Currency Deposits/GDP Retail Lending Margin in Foreign Currency Foreign Currency Loans/GDP.3.3 Foreign Currency Deposits/GDP =GoodsMarketMoneyDemand,---=AssetMarketMoneyDemand

24 Contraction in Bank Money Supply Comparison: Different Substitutabilities between Domestic and Foreign Deposits (magnitudes: 1%ofGDP ; θ d =1(UIP)/1.5/.5) -.1 GDP Base Alt1 Alt Home Inflation.5 -. Real Policy Rate Foreign Inflation Depreciation Real Exchange Rate Consumption Exports/GDP Imports/GDP Terms of Goods Trade 1. Share of Foreign Goods in Consumption.3 Current Account/GDP Terms of Money Trade (RR_D_H/RR_D_F) Share of Foreign Ccy Deposits in Total Deposits Price of Land Retail Lending Margin in Home Currency.8.8 Home Currency Loans/GDP Home Currency Deposits/GDP Retail Lending Margin in Foreign Currency Foreign Currency Loans/GDP Foreign Currency Deposits/GDP.6.4. =UIP,---=IntermediateSubstitutability,... =LowSubstitutability 3

25 Contraction in Bank Money Supply 3 Comparison: Different Inflation Response of the Policy Rate (magnitudes: 1%ofGDP ; θ d =.5 ; m i /m π =.7/. /.85/1.5 /.95/1.5) GDP Base Alt1 Alt Home Inflation Real Policy Rate 3 Foreign Inflation 3 Depreciation Real Exchange Rate Consumption Exports/GDP Imports/GDP Terms of Goods Trade Share of Foreign Goods in Consumption Current Account/GDP Terms of Money Trade (RR_D_H/RR_D_F) Share of Foreign Ccy Deposits in Total Deposits Price of Land -6-6 Retail Lending Margin in Home Currency.8.8 Home Currency Loans/GDP Home Currency Deposits/GDP Retail Lending Margin in Foreign Currency Foreign Currency Loans/GDP Foreign Currency Deposits/GDP =UIP,---=IntermediateSubstitutability,... =LowSubstitutability 4

26 Flight to Safety: Money Demand Increase by Home Households Comparison: Goods Market versus Asset Market Money Demand (magnitude: 1% of GDP) -1 GDP Base Alt1 Alt -1 1 Home Inflation Real Policy Rate Foreign Inflation Depreciation 1 Real Exchange Rate Consumption Exports/GDP Imports/GDP Terms of Goods Trade 3 Share of Foreign Goods in Consumption Current Account/GDP Terms of Money Trade (RR_D_H/RR_D_F) Share of Foreign Ccy Deposits in Total Deposits Price of Land Retail Lending Margin in Home Currency Retail Lending Margin in Foreign Currency Home Currency Loans/GDP Foreign Currency Loans/GDP Home Currency Deposits/GDP Foreign Currency Deposits/GDP =GoodsMarketMoneyDemand,---=AssetMarketMoneyDemand 5

27 Flight to Dollars: Dollar Demand Increase by All Households Comparison: Goods Market versus Asset Market Money Demand -.1 GDP Base Alt1 Alt Home Inflation Real Policy Rate Foreign Inflation Depreciation Real Exchange Rate Consumption Exports/GDP Imports/GDP Terms of Goods Trade Share of Foreign Goods in Consumption Current Account/GDP Terms of Money Trade (RR_D_H/RR_D_F) Share of Foreign Ccy Deposits in Total Deposits Price of Land Retail Lending Margin in Home Currency.6 Home Currency Loans/GDP.6.5 Home Currency Deposits/GDP Retail Lending Margin in Foreign Currency Foreign Currency Loans/GDP Foreign Currency Deposits/GDP =GoodsMarketMoneyDemand,---=AssetMarketMoneyDemand 6

28 5 Conclusions The Monetary Approach to the Exchange Rate predates modern modeling. But it incorporates many insights that we might be ignoring at our peril. The most important of these is the importance of money. However: It focuses only on the small and endogenous high-powered money. 97% of money used today is instead produced through bank lending. In this paper we therefore fuse two approaches: 1. The Monetary Approach to the Exchange Rate.. The Financing (rather than Intermediation) Approach to Banking.

29 The Financing Approach has a focus on gross positions at its center: Increasingly recognized as critical in Open Economy Macro. Economic decisions are always on gross positions. Net positions are accounting residuals. First results are very interesting: Shocks to supply and demand of money. Bank loans can instantaneously respond to such shocks. The response is through the creation of ledger entries. Real saving does not constrain their response in the least. Strong real effects via liquidity shortages/excesses. Effects on the exchange rate via a modified UIP condition.