Purchasing Power Parity, Wages and Inflation in Emerging Markets

Article
Purchasing Power Parity,
Foreign Trade Review
49(4) 327–347
Wages and Inflation in
©2014 Indian Institute of
Foreign Trade
Emerging Markets
SAGE Publications
Los Angeles, London,
New Delhi, Singapore,
Washington DC
DOI: 10.1177/0015732514543587
http://ftr.sagepub.com
Ashima Goyal
Abstract
Three open economy puzzles are, first, the price level of an advanced economy
(AE) is higher compared to that of an emerging or developing economy (EDE)
measured in the same currency. But, second, inflation tends to be much higher
in EDEs. Third, there is a persistent deviation of real exchange rates from
equilibrium values although nominal shocks, which cause such deviation, are
expected to have only short-run effects. Balassa Samuelson (BS) explained the
first puzzle by introducing two sectors with differential productivity growth.
The article presents a framework that allows a new simple proof of the BS
result. Inflation in EDEs is often attributed to the catch-up process where pro-
ductivity in traded goods starts growing faster than that in non-traded goods.
But if wages grow at a relatively higher non-traded goods productivity growth,
nominal shocks can trigger deviations from equilibrium and adjustments that
sustain inflation. This explanation retains the BS ranking of relatively lower
traded goods productivity in EDEs, yet explains the second and third puzzles.
Tests with Indian data support the hypotheses, suggesting that they provide a
better explanation of Indian inflation.
JEL: F41, E31
Keywords
Purchasing power parity, inflation differentials, wage rigidities, emerging markets
Introduction
The article addresses three puzzles. First, the price level of an advanced economy
(AE) is higher compared to that of an emerging or developing economy (EDE)
measured in the same currency. Yet, inflation tends to be much higher in EDEs.
This is the second puzzle.
Ashima Goyal, Professor, Indira Gandhi Institute of Development Research, Mumbai,
India. E-mail: ashima@igidr.ac.in

---

328
Ashima Goyal
The first indicates a failure of trade arbitrage, or purchasing power parity
(PPP), which should cause prices of the same goods to converge across countries
when measured in the same currency. Deviations of the real exchange rate from
its PPP value of unity are common and persistent. Monetary and financial factors
explain short-term volatility of nominal exchange rates. But since monetary fac-
tors are not expected to have long-term effects, the half-life of 3–5 years for devi-
ations from PPP is the third puzzle. Rogoff (1996) comes to the conclusion that
the answer must lie in factors such as transactions costs and other barriers that
prevent perfect trade integration. But, this is an ad hoc explanation. Moreover,
while the degree of integration varies with time and the tariff regime in place,
deviation from PPP has been a constant feature. Recent research, that applies
sophisticated time series tests, and allows for non-linear adjustment, has reduced
the degree of deviation, but it is still there.1
A less stringent version of PPP, relative PPP (RPPP), is also not empirically
supported. RPPP is consistent with persistent deviation in price levels across
countries, as long as nominal depreciation equals the inflation differential so that
the real exchange rate does not change. But persistent changes in the real exchange
rate do occur. Therefore, it is worthwhile to explore if monetary shocks can affect
the real exchange rate in the longer run.
An explanation for the first puzzle, from Balassa (1964) and Samuelson (1964)
(BS), is that although absolute productivity of AEs is higher for all goods, the
productivity gap is greater in traded goods. If wages are equalized across all sec-
tors in each country, the price of non-traded goods such as services would be
higher in AEs, pushing up the price level and the real exchange rate.2
We argue that large nominal shocks and structural features that make for their
persistence can explain the puzzles in the BS framework that distinguishes
between traded and non-traded goods. In the PPP theories, real factors—trade
arbitrage—determine the real exchange rate. When nominal exchange rate shocks
interact with aspects of structure, such as wage setting behaviour, there can be
persistent effects, including on real exchange rates and on inflation, which explain
the second and third puzzles. Starting from the 1970s, capital flows to EDEs
became large and volatile. These drove changes in the nominal exchange rate,
independently of money supply shocks, as many EDEs moved towards more flex-
ible exchange rate regimes.
We prove the standard BS result in a simple framework with equal rates of
wage growth in both traded and non-traded sectors and then extend it by dropping
the assumption that the price level is fixed in the EDE. An extension allows for
shocks such as nominal depreciation and rise in nominal wages and explains the
higher EDE inflation. It shows how a real wage target in terms of food prices can
sustain inflation as nominal wages rise in response to a nominal depreciation. This
differs from the standard result where causality runs from relative money sup-
ply growth to nominal depreciation. Under a floating exchange rate regime,
capital flows can change the nominal exchange rate without prior change in
Foreign Trade Review, 49, 4 (2014): 327–347

---

Purchasing Power Parity, Wages and Inflation in Emerging Markets
329
money supply, so that the nominal depreciation from an outflow is better analyzed
as an exogenous shock. Then, large risk-on risk-off capital flows that drive nomi-
nal exchange rates and interact with different types of wage rigidities can explain
bouts of inflation in EDEs. This reverses the standard explanation that excess
money supply causes both inflation and nominal depreciation.
In order to focus on BS type productivity differentials as driving relative prices,
we assume simple linear production technologies. So, we abstract from cyclical
effects. Unemployment is analyzed when we turn to the short term, but in our
simplified framework, it is driven entirely by adjustments to achieve external
balance.
With catch-up growth, according to the standard BS result, prices rise because
wages rise at the higher traded goods productivity growth. So, emerging market
and developing economies (EMDE) inflation can be explained by its transition to
an AE. But our framework suggests that nominal and real depreciation can drive
EMDE inflation even though BS sectoral productivity rankings remain unchanged,
with non-traded good prices higher in AEs. If wages grow at a relatively higher
non-traded goods productivity growth, nominal shocks can trigger deviations
from equilibrium and adjustments that sustain inflation. This explains the second
puzzle, higher EDE inflation, as well as the third, sustained deviations from PPP
exchange rates.
Evidence from India supports the hypotheses, suggesting that they provide a
better explanation of Indian inflation. Following the exogenous international food
price shocks in 2007, India had high food price and wage inflation. Exchange rate
volatility was also high as a consequence of risk-on risk-off capital flows follow-
ing the global financial crisis (GFC). We exploit state-level wage data and the
exogenous shocks that cut through simultaneity issues, to test our hypotheses that
nominal depreciation affects inflation through its impact on wages. The results
show depreciation and food price inflation significantly and positively affect
wages, thus supporting the role of external shocks and of wage behaviour as
explanations of inflation and validating the extended BS results derived. Although
government employment insurance schemes were widely blamed for wage infla-
tion, convergence across states was limited. This should have occurred if central
schemes had a major impact. Aggregate demand variables also were not signifi-
cant as explanatory variables.
Policy implications are drawn out. For example, strategic use of exchange rate
policy can abort a wage-price cycle in response to an external shock. Since mon-
etary shocks have persistent real effects in the presence of real wage rigidities,
macroeconomic policies have a critical role.
The remainder of the article is as follows: After labour market equilibrium
conditions presented in the second section, the third section derives the basic PPP
and BS hypotheses. The fourth section explores wage-price-exchange rate dynam-
ics, the fifth section tests the hypotheses with Indian data and the sixth section
draws out implications for policy, before the last section concludes.
Foreign Trade Review, 49, 4 (2014): 327–347

---

330
Ashima Goyal
The Model
Consider an EDE with two sectors producing traded and non-traded goods.
Production comprises manufacturing and agriculture (A) and services (N). We
assume simple linear production functions of the form Y = aL,3 where a is productiv-
ity. Manufacturing and agriculture are traded goods. Services are non-traded. Modern
manufacturing tends to be a high skill and wage enclave delinked from the much
larger informal economy consisting of agriculture and services. We focus on this
more integrated informal economy. If profit maximization equates nominal wages to
the marginal product of labour (MPL) in each of the A and N sectors, then:

W = P MPL = P a
(1)
A
A
A
A A

W = P MPL = P a ...