On the Implications of Structural Transformation for Inflation and Monetary Policy (Work in Progress)

Size: px

Start display at page:

Download "On the Implications of Structural Transformation for Inflation and Monetary Policy (Work in Progress)"

Sophia Wright
5 years ago
Views:

1 On the Implications of Structural Transformation for Inflation and Monetary Policy (Work in Progress) Rafael Portillo and Luis Felipe Zanna IMF Workshop on Fiscal and Monetary Policy in Low Income Countries International Growth Centre November 2-3, 212

2 Motivation Many SSA countries re-anchored inflation in the early 2s. In the context of fiscal based stabilization efforts. Money targeting, to signal stabilization was on track. Inflation has remained volatile. Some countries are moving toward Inflation Targeting. The inflation forecast is the intermediate target. Focus on understanding why inflation deviates from target, determining what policy decisions are necessary (if any) to bring it back in line. What should we expect from inflation in LICs as these countries adopt more modern monetary policy frameworks?

3 Motivation LICs find themselves at earlier stages of structural transformation. Low income per capita (IPC). Share of spending on agricultural products (ω F ) is large. Income and price elasticities for food are lower. Food shocks should have large effects on food prices (p F ). Flexible food prices: p F will also capture any incipient aggregate demand pressures. These features have implications for inflation (π): π = π N + ω F (π F π N ) = π N + ω F p F Holding σ πn constant, inflation in LICs should be: (i) volatile, and (ii) dominated by changes in the relative price of food (ω F p F ).

4 In this paper We present a simple model of structural transformation to quantify the importance of these features for inflation and monetary policy. Main features: Two sectors: food and non food. Stone Geary preferences: ω F changes with aggregate TFP. Food prices are flexible, non food prices are sticky. Shocks to food productivity (e tf ) and monetary policy (e MP ). MP: interest rate rule s.t. e tf does not affect π N. The model encompasses the US and Africa (ω F, IPC). (σ etf,σ emp ) s.t. the model reproduces (σ π,σ πn,σ pf ) for US (US serves as benchmark).

5 Main findings 1. Structural features amplify the effects of food shocks on p F and on π at earlier stages of economic development. Effect on output is larger (at earlier stages of development). 2. Structural features also amplify the effects of shocks to monetary policy on inflation at earlier stages of development. Effect on output is similar.

6 Main findings(2) 3. Inflation in the median African country should be 57 percent more volatile than US inflation ( p F percent). Falls short: quarterly inflation (1995:211) is 3 percent more volatile than US ( p F is 2 percent) percent of inflation volatility in the median African country should be accounted by ω p F (3 percent in the US). Matches the relative decomposition in the data. 5. The larger the food share (sector), the more costly it is to target headline inflation (in terms of the total output gap). Greater price flexibility (in the economy) is associated with larger real effects of alternative nominal anchors.

7 Related work Structural transformation: Caselli and Coleman (21), Kongsamut et al (21), Ngai and Pissarides (27), Rogerson (28), among others. Food prices in inflation in emerging markets and low income countries: Catao and Chang (21), Anand and Prasad (21), Walsh (21), IMF WEO (211), Portillo and Zanna (212), Adam et al (212), Andrle et al (212). Choice of inflation target: Aoki (21),...

8 Stylized facts: Food Share in CPI against GDP per Capita Figure: Blue diamonds: countries; dashed red line: log trend; dashed blue line: model.

9 Stylized facts: Inflation volatility against GDP per capita Figure: Blue diamonds: countries.

10 Stylized facts: Volatility of p F against GDP per capita Figure: Blue diamonds: countries.

11 The model: Consumers Consumers maximize lifetime utility: ( ) E β t logct n1+ψ t 1 + ψ where c is given by: t= ct = D (c F,t c F ) α F c 1 α F N,t subject to the budget constraint: P F,t c F,t + P N,t c N,t + B t+1 = W t n t + Π t + R t 1 B t.

12 The model: Consumers First order conditions: n ψ t = w t c 1 t ( ) 1 cf,t = c PF,t F + α F Pt ct c N,t = (1 α F ) ( PN,t P t ) 1 c t c 1 t = βe t [c 1 t+1 R t π t+1 ] wt = W t Pt, Pt = P α F F,t P1 α F N,t, πt = P t Pt 1

13 The model: Food and non food sector Food: perfectly competitive, flexible prices. Shock to sectoral productivity t t. Non food: monopolistic competitition, sticky prices a la Calvo (1983), employment subsidy. Production function: y i = At t A i ni,t α, for i = N, F. A: economy wide total factor productivity.

14 Steady State Steady state structure of the economy varies with aggregate TFP (A). P N = P F = P. w = A. c = f (A). c F = c F + α F c, c N = (1 α F )c. Actual consumption c = c F + c N = c F + c. ω F = c F +α F c c F +c α F, as A. Income and Price elasticities for food: Price: α F ω F. Income: α F ω F.

15 Log-linearized version of the model Consumers: where ŵ t = Ŵt ˆP t. ψˆn t = ŵ t + (ω F α F )ˆp F,t 1 α F 1 ω F ŷ t, ŷ t = ŷ t+1 1 ω F 1 α F (ˆR t ˆπ t+1 + (ω F α F ) ˆp F,t+1 ), ˆπ t = ˆP t ˆP t 1 = ˆπ N,t + ω F ˆp F,t. ŷ F,t = α F ω F ŷ t α F ω F (1 ω F )ˆp F,t ŷ N,t = 1 α F 1 ω F ŷ t + α F ˆp F,t

16 Log-linearized version of the model Supply side: ŷ t = ω F ŷ F,t + (1 ω F )ŷ N,t, ŵ t (1 ω F )ˆp F,t = 1 α α ŷf,t + αˆt 1 t ˆπ N,t = βˆπ N,t+1 + κ 1 ω F ( 1 α α ŷt + ŵ t ω ) F α ˆt t.

17 Log-linearized version of the model Monetary Policy ˆR t = (ˆr n t + ω F ˆp F,t+1 ) + ϕˆπ N,t + dev t Shock to food production dev t = ρ dev dev t 1 + e MP ˆt t = ρ ˆt t 1 ξˆt t 2 + e tf

18 Calibration Parameter Value c F.99 α F.71 α.7 κ.858 ϕ 1.5 ψ 5 ρ.8 ρ d ev.8 ξ.1 σ MP.5 σ tf.55

19 Simulations

20 e MP < Deviation from monetary policy rule Change in relative price of food Non food inflation Headline inflation Output Nominal interest rate Poor country Rich country Food production Non food production

21 e tf < Productivity in the food sector Change in relative price of food Non food inflation Headline inflation Output Natural rate of interest -2-1 Poor country Rich country Food production Non food production

22 Second Order Moments Data Model σ(tf) =.55 σ(dev) =.5 US Med. Afr Ratio Rich Poor Ratio Rich Poor Ratio Food share in CPI (percentage) GDP per capita % % PPP (thousands of dollars, avg 21-21) Volatility (std, BP filtered quarterly data, 1995:I-211:IV)) Headline inflation (.3,.6) (.23,.45) (.24,.47) (.32.62) Non Food inflation (.22,.46) (.21,.43) Changes in the relative price of food (.37,.73) (.54, 1.7) (.4,.82) (.41,.84)

23 Second Order Moments Data Model σ(tf) =.55, σ(dev) =.5 US Med. Afr Ratio Rich Poor Ratio Food share in cpi GDP per capita % % PPP (thousands of dollars, avg 21-21) Volatility (std deviations, BP filtered quarterly data, 1995:I-211:IV)) Headline inflation (.25,.49) (.37,.66) Non Food inflation (.22,.46) (.21,.43) Changes in the relative price of food (.54, 1.3) (.67, 1.3) Inflation decomposition Non food food weight* changes in the relative pri

24 ˆπ t = Productivity in the food sector Change in relative price of food Non food inflation Headline inflation Output Policy Rate -2-1 Poor country Rich country Food production Non food production

25 Extensions and future work Incorporate other aspects of structural transformation in the model and assess their quantitative contribution to inflation and relative price volatility. Monetary policy design in LICs in the context of various food shocks. What economic features are necessary to make the volatility of the relative price of food an appropriate objective of monetary policy? Joint work with Chris Adam and Steve O Connell. Features: Subsistence requirements in food consumption Limited financial participation Real wage rigidities in the non food sector Costly internal transport

26 5. Under Stone-Geary preferences, two distinct price levels emerge. The measured price level P, with food weight ω F. A notional price level P, with weight α F. ω F α F as TFP, ω F > α F. P, not P, is what matters for private sector decisions. α F /ω F helps define price and income elasticities.

27 The model: Food and non food sector Food: perfectly competitive, flexible prices. Production is given by: y F,t = At t A F n α F,t Non food: a continuum of producers, each producing a variety y(i) N,t. Demand is given by y(i) N,t = y N,t ( P(i) N,t P N,t ) ɛ, with P N,t = [ P(i) 1 ɛ N,t ] 1 1 ɛ. Production is given by: y(i) F,t = AA N n(i) α N,t Firms face a random signal that allows them to change prices with probability θ, a la Calvo (1983). Firms also receive an employment subsidy τ.

The Risky Steady State and the Interest Rate Lower Bound

The Risky Steady State and the Interest Rate Lower Bound Timothy Hills Taisuke Nakata Sebastian Schmidt New York University Federal Reserve Board European Central Bank 1 September 2016 1 The views expressed