Mobile Money and Monetary Policy

Transcription

1 Mobile Money and Monetary Policy Christopher Adam and Sébastien Walker University of Oxford 12 February 2015

2 Outline Motivation: Mobile Money and Monetary Policy An alternative framework: Anand and Prasad (2010) Mobile money in the Anand and Prasad model Policy implications

3 Punchlines Financial frictions supply shocks favour targeting headline vs. core ination, especially with mostly rural population Mobile money reduced volatility of all important variables Gains mostly to rural households

4 Is mobile money a threat to the conduct of monetary policy? The Implications of Innovations in the Financial Sector on the Conduct of Monetary Policy in East Africa David Weil, Isaac Mbiti, Francis Mwega (IGC WP, 2013). By 2011: 70% of households use regularly use mobile money 18m registered users (compared to 1.4m with ATM cards)...we conclude that the monetary implications of mobile money are currently minimal in Kenya. However...the developments and innovations in this space could fuel the growth of mobile money such that it reaches levels where it could have implications for monetary policy

5 The demise of money-based frameworks? Eective reserve-money targeting as practiced in East Africa relies on: predictability in the velocity of circulation of broad money (private sector demand behaviour) predictability in the money multiplier (the policy control / transmission mechanism) Instability simple relationship between money and ination dicult to predict. Financial liberalization and innovation end of money targeting in OECD... Is it doing the same in East Africa?

6 Mobile money a threat to the conduct of monetary policy? Central banks have expressed concern that these new technologies undermine the ecacy of monetary policy as conventionally conducted. Kenya Monetary Ratios M-Pesa Launched July 2007 Currency/Deposit Ratio [Left Scale] M2 money multiplier [Right Scale]

7 An East African transition towards IT frameworks East African Countries moving rapidly to IT-based frameworks Uganda, formal committment to `IT-Lite' in July 2011 with target of 5% for core ination Kenya, formal committment to an ination target of 5% (+/- 2.5%) for headline ination In other EAC countries, committment to keep headline ination in stable single digits All central banks actively moving towards frameworks that can steer short-term market rates into a closer relationship with the policy interest rates (`bank rate')

8 The conventional wisdom Demand shock: ination, output (or vice versa)' Supply shock: ination, output (or vice versa). If supply shocks dominate strict IT exacerbates output volatility

9 Implications for LICs Conventional IT solution: target core ination; accomodate non-core price shocks (but not second-round eects) But if these are frequent and large an IT regime targeting core ination allows high volatility in headline ination. General result: the broader the measure of ination you seek to stabilize, the greater will be output volatility... Other (i.e. scal) policy instruments to stabilize output?

10 Why supply shocks are likely to dominate in LICs The Food Engel curve TZA Share of food in total consumption GDP per capita at constant 2001 international prices Source: USDA Economic Research Service Christopher Adam. 1

11 Supply shocks dominate in LICs Correlation between output gap and ination: demand shocks ρ > 0, supply shocks: ρ < 0 Correlation between real GDP and inflation TZA Real GDP per capita, PPP HP-filtered variables. All countries with 17 or more yrs of data starting in Source: Adam et al. (2010). Christopher Adam. 1

12 Anand and Prasad (2010): countering the conventional wisdom for emerging market economies Optimal Price Indices for Targeting Ination Under Incomplete Markets Rahul Anand and Eswar Prasad (IMF WP 20/200) Adapt conventional New-Keynesian model to dualistic setting: `Food-producing'/rural households: hand-to-mouth/keynesian consumers `Non-food-producing'/urban households can borrow/save Monetary policy transmission through consumption Euler equation for urban households.

13 Anand and Prasad(2010): key results and intuition With incomplete markets, targeting core ination may not be optimal. Flexible headline ination targeting generally dominates. Absence of nancial markets means that relative price shocks in the food (ex-price) sector have direct income eects for credit constrained households => aggregate demand eect which does not respond to conventional demand-side policy responses. Our research question: how robust are these ndings when we allow for mobile money technology in this class of model?

14 Adding mobile money transfers to Anand and Prasad (2010) FOOD FIRMS NON-FOOD FIRMS WAGES LABOUR WAGES LABOUR FOOD HHs FOOD NON-FOOD NON-FOOD HHs REMITTANCES/ MOBILE MONEY TRANSFER INTEREST RATE CREDIT/SAVINGS CENTRAL BANK BOND MARKET

15 Urban households Urban HHs' demand for composite consumption good (C s t ): [ ] (C (C s t ) σ ) s σ R t = βe t t+1 Π t (1) β: discount factor R t : Gross nominal interest rate Π t : Gross ination rate σ: inverse of elasticity of intertemporal substitution

16 Rural households Rural HHs' demand for composite consumption good (C f t ): C f t = x f,ty f,t x f,tc + m t 1 + µ (2) x f,t: relative price of food y f,t: food production C : subsistence consumption level m t : remittances µ: remittance `melt' rate

17 Remittances Remittance payment: m t = me κ ( Ωt Ω 1 ) (3) m t : remittances m: steady-state m t Ω t : rural HHs' pre-remittance income net of subsistence consumption Ω: steady-state Ω t κ: elasticity of m t w.r.t. Ω t

18 Interest rate rule log ( Rt ) = ρ i log R ( Rt 1 R ) + ρ π log ( Πt Π ) + ρ y log ( Yt Y ) (4) R t : Gross nominal interest rate Π t : Gross headline or core ination rate Y t : GDP ρ i = 0.7, ρ π = 2, ρ y = 1

19 Experiments 1. Targeting headline ination vs. core ination 2. Three remittance set-ups: `No remittances': remittances xed at steady-state level 1 `Constrained remittances': elasticity w.r.t. rural incomes, 2 20% `melt' `Mobile money': unit elasticity w.r.t. rural incomes, no `melt'

20 Proportional change in std. deviations: headline- over core-ination-targeting, food shock NR = no remittances, CR = constrained remittances, MM = mobile money Remittance set-up NR CR MM Headline ination -29% -20% -13% Core ination -32% -14% 2% GDP -32% -8% 84% Nominal int. rate 48% 93% 218% Rural cons. -9% -5% -3% Urban cons. 4% 8% 16%

21 Proportional change in std. deviations: CR/MM over NR, food shock NR = no remittances, CR = constrained remittances, MM = mobile money Ination target Headline Core Remittance set-up CR MM CR MM Headline ination -16% -32% -25% -44% Core ination -10% -21% -29% -47% GDP -30% -51% -48% -82% Nominal int. rate -25% -48% -43% -76% Rural cons. -37% -73% -39% -74% Urban cons. -34% -64% -37% -68%

22 Std. deviations by regime 10, food shock NR = no remittances, CR = constrained remittances, MM = mobile money Ination target Headline Core Remittance set-up NR CR MM NR CR MM Headline ination Core ination GDP Nominal int. rate Rural cons Urban cons

23 Std. deviations by regime 100, non-food shock NR = no remittances, CR = constrained remittances, MM = mobile money Ination target Headline Core Remittance set-up NR CR MM NR CR MM Headline ination Core ination GDP Nominal int. rate Rural cons Urban cons

24 Policy implications No threat to eective conduct of monetary policy with move from reserve-money targeting to IT Mobile money use macroeconomic stability, encourage further spread A possible case for targeting core ination

25 Punchlines Financial frictions supply shocks favour targeting headline vs. core ination, especially with mostly rural population Mobile money reduced volatility of all important variables Gains mostly to rural households