Exchange Rate Pass-Through in Developing and Emerging Markets: A Survey of Conceptual, Methodological and Policy Issues, and Selected Empirical Findings

Figures & data

Table 1. Typology of multi-country studies for selected emerging market countries investigating ERPT

Study	Sample and structural break-points	Country coverage	Single equation or System†	Controls in equations	Is long-run price homogeneity satisfied?	Which ERPT results are reported?
Consumer price aggregate (CPI)
Razafimahefa (2012) Panel and time series Quarterly CPI (IFS) and NEER (BIS)	Sample: 1985Q1–2008Q2.Breakpoint: Chosen for the largest fall in ERPT coefficient as 1997 (apparently all four quarters). The ERPT results are also averaged and analysed by fixed/flexible regime groups.	SSA as a whole (panel); and individual countries (unspecified countries dropped due to lack of data).	SYSTEM for panel Differenced VAR models, 1 or 2 lags:a. bivariate with only the exchange rate and the CPI (time series and balanced panel);b. above, plus levels term with CPI and NEER (details unclear). (Balanced panel only, fixed effects, OLS).SYSTEM for individual countries Lag lengths not reported (Schwartz criterion).	No destination cost controls. No exporter cost controls. No demand controls.SSA panel: examine also directional asymmetry with depreciation dummy and fall in ERPT with post-1997 dummy (interacting with Δ log NEER).	No domestic or foreign prices are included: long-run homogeneity is violated (in specif. b., only CPI and NEER are assumed to cointegrate).	1-, 4- and 8-quarter ERPT for individual countries from accumulated normalised impulse responses.For the SSA panel, ‘long-run’ ERPT (see online Appendix 1 for criticism). Correlations given between unspecified ERPT elasticities and macro/micro drivers.
Kohlscheen (2010) Time series Monthly CPI (IFS) and NEER (BIS)	Sample: from inception of the floating exchange rate regime ‘as defined by Reinhart and Rogoff (2004)’, until October 2008. Breakpoint: No breakpoint. Exchange rate regimes are isolated as above.	Brazil, S. Korea, Mexico, Indonesia, S. Africa, Thailand, Czech Republic, Philippines.	SYSTEM Differenced VAR models, 12 lags:a. bivariate with only the exchange rate and the CPI;b. above, plus the (differenced) log of output. Tests robustness to ordering.	No destination cost controls. No exporter cost controls. Destination demand proxied by the log of the industrial production index.	No long-run terms.	3-, 6-, 9-, 12-month ERPT for individual countries from accumulated impulse responses. Simple correlations between 6- and 12-month ERPT elasticities and macro/micro drivers.
Akofio-Sowah (2009) Panel Quarterly CPI (IFS) and NEER (IFS)	Sample: 1980–2005 maximum; samples are unclear but short; SSA region: 332 observations or 5 years data per country on average; LA region: 189 observations or 4 years data per country on average.Breakpoint: No breakpoint. Exchange rate regime groups identified by dummies: 4 currency unions (Africa); 3 dollarisation groups (LA).	15 sub-Saharan African and 12 Latin American countries (online Appendix 2); grouped as SSA/LA. (Removed 4 hyper-inflation countries).	SINGLE EQUATION Differenced equation:a. current Δlog (NEER); one lag in the dependent variable; currency group dummies and current control variables (see RHS).(Unbalanced panel with fixed effects, separately estimated for LA and SSA regions).	No destination cost controls. Exporter costs proxied by trade-weighted CPI.The currency group dummies, real output growth, change in trade openness, change in exchange rate volatility each interacted with current Δlog (NEER). (Last 3 not included/tested separately).	No long-run terms.	Reports the regressions for SSA and LA. Does not give the net ERPT after one quarter by adjusting for the interactive terms (footnote 2). But no interactive terms are significant except change in volatility for SSA, and currency groups for both SSA and LA.Coefficients on currency groups show the relative ERPT of groups.
Mihaljek and Klau (2001, 2008) Time series Quarterly CPI (IFS) and NEER (IFS)	Sample: 1994Q1 to 2006Q2 (2008 paper) and (max. sample) 1981–2001 (2001 paper).Breakpoint: ERPT coefficients are compared for the above two overlapping samples. However, specifications differ for the 2 papers; moreover in 8/14 countries the samples only begin in the mid-90s (for example, Turkish sample begins in 1995Q2 in earlier paper).	14 emerging market countries (online Appendix 2).	SINGLE EQUATION Differenced equations:a. current Δlog (NEER) and up to 3 lags; up to 3 lags on dependent variable; and current and up to 3 lags on the control variables.b. the above, plus current and up to 2 lags of REER gap. (from 2008 paper).	No destination cost controls. Exporter costs proxied by trade-weighted CPI. Destination demand proxied by output gap. Regressions run with and without the REER gap: effectively introduces a weighted foreign CPI term in the long-run.Size and directional asymmetry tested with simple dummies (that is, not interacted with Δlog (NEER)).	REER gap is included: the REER minus a long-run trend (calculated using the HP filter). This imposes homogeneity (PPP); and the trends potentially capture evolving productivity differentials amongst countries.	4 quarter ERPT (‘short run’) for individual countries in Graph 5, with and without controlling for long-run with REER gap (from their Tables A2 and A3). The ERPT estimates are incorrect as they fail to correct for REER gap and lagged dependent variable (see footnote 2).The stability of ERPT result is unreliable (see column  2 at left).
Choudhri and Hakura (2006) Time series Quarterly CPI (IFS) and NEER (IFS)	Sample: 1979–2000. Breakpoint: Samples are divided by inflation regimes (their Table 3) and regressions are run separately for each.	71 countries including developing countries.	SINGLE EQUATION Differenced equation: a. current & Δlog (NEER) and lags; several lags in the dependent variable; and several lags in the control variable (see RHS) beginning at (t–1). Lag order determined by Schwartz criterion – not reported.	No destination cost controls. Exporter costs proxied by trade-weighted CPI. No demand controls.	No long-run terms. Did not find evidence for a stationary REER using country and panel tests.	0-, 1-, 4- and 20-quarter ERPT for all countries with single or dual inflation regimes (the lagged dependent variables are dealt with as in footnote 2). Cross-section regressions of 0- and 20-quarter ERPT elasticities on macro drivers and openness.
Goldfajn and Werlang (2000) Panel Monthly CPI (IFS) and NEER (IFS)	Sample: the maximum period: 1980–1998 (unclear if it is a balanced panel). Breakpoint:No breakpoints. Countries are grouped by indebtedness. Interaction terms with trade openness are included.	71 countries (their Table A1) in groups: by region; developed, developing (28) and emerging (24); more or less indebted; OECD and non-OECD.	SINGLE EQUATIONDifferenced equation: a. one lag in dependent variable at (t–1); lagged control variables (see RHS) all at time (t–1); the dependent variable (inflation rate) and regressor exchange rate depreciation are both accumulated for j periods ahead: inflation [t, t + j], and Δlog (NEER) [t–1, t + j–1]; that is, the latter first included at (t–1).(There are overlapping data.)(Panel with fixed effects, GLS.)	No destination cost controls. No explicit exporter controls, though the REER gap is included: effectively introduces a weighted foreign CPI in the long run. Demand controlled for by real GDP (or industrial prod index) gap. Controls include openness and sometimes survey expectations of exchange rates. The GDP gap, inflation, openness and REER gap are each interacted with Δlog (NEER) [t–1, t + j–1].	REER gap is included: the REER minus a long-run trend (calculated using the HP filter). This imposes homogeneity (PPP) and the trends potentially capture evolving productivity differentials amongst countries.	1-, 3-, 6-, 9-, 12-, (18-) month ERPT for all countries, for regions and for groups, with and without the interaction terms.
Consumer price aggregate (CPI) and import prices disaggregated to products (both at the dock and retail)
Frankel, Parsley and Wei (2012) Panel Annual IUV of 8 commodities ‘retail’ imports (EIU, footnote 4)‡; matching products IUV ‘at the dock’ imports (UN); aggregate CPI (WEO); and bilateral nominal exchange rates (EIU).	Sample: 1990–2002. Breakpoint: No specific breakpoint. Interact each of the ecm term and Δlog (NEER) with a trend, for the sample as a whole.	One group of 76 countries; and sub-grouped: developing/high income.	SINGLE EQUATION Differenced equation: a. base model: current Δlog (NEER); no lags in dependent variable; current exporter cost (see RHS) and ecm (at t–1); b. the above, plus interaction terms for each of the ecm term and Δlog (NEER): with macro/micro drivers and a trend (see RHS).(Pools the 8 goods, 12years and 76 countries with product and/or country dummies)	No destination cost controls in base model. Exporter costs: matching prices from exporting countries for 8 import goods in import equations; in CPI eqn, uses trade-weighted foreign CPI. Interaction terms in models b with: linear trend, log measures of: income, tariff, distance, size, destination real wage, average inflation; and exchange rate volatility. (The uninteracted variables were not tested separately in regression.)	Long-run PPP is imposed in ecm term; that is, homogeneity for foreign prices is satisfied.	Base model a. gives ERPT after one year for all countries, and developing/high income sub-groups. Net ERPT not calculated for models b. (see footnote 2)For 3 dependent variables CPI, ‘at the dock’ and ‘retail’ specific import product prices: in the models b. the interaction terms are added sequentially (that is, specific-to-general). Variables in the interaction terms were not de-meaned (see footnote 2).
Import price aggregate
Brun-Aguerre, Fuertes, and Phylaktis (2012)**Panel and time series Quarterly IUV and NEER (see their WP version)	Sample: for time series and unbalanced panel the maximum sample is 1980Q1–2009Q3; balanced panel sample is 1997Q1–2009Q3; for drivers of ERPT, sample is 2004Q3–2009Q3.Breakpoint: No breakpoint.	18 emerging markets and 19 developed markets (Online Appendix 2).	SINGLE EQUATION Three differenced specifications: a. current Δlog (NEER); one lag in dependent variable; current exporter cost (see RHS) and ecm (at t-1); b. above, plus 4 lags on Δlog (NEER) and on exporter cost but no lagged dependent variable or ecm; c. as in b. but with ecm (at t–1). (Equations estimated for each country, OLS.)	No destination cost controls. Exporter costs use a contentious proxy – see Section 5: the import-weighted total export unit value indices of trading partners. No demand controls. Panel analysis of ERPT drivers examines: foreign exchange volatility; asymmetry of ERPT; import dependence; tariffs; inflation; relative wealth; global PMI index; output gap.	Freely-estimated ecm at (t–1) in models a. and c.: a linear combination of levels of IUV, foreign price and exchange rate. Long-run homogeneity for foreign prices is not imposed.	For each of the models a., b. and c., 1-quarter (short-run) ERPT for individual countries, and averaged for all, over (i) full sample (unbalanced panel), and (ii) balanced panel sample. Long-run (equilibrium) ERPT from ecm for models a. and c. For model b., five-quarter ERPT is reported.Panel (fixed effect) regressions of short term ERPT elasticities on macro/micro drivers.
Barhoumi (2006) Time series (also run panel regressions to test for homogeneous long-run response across countries) Annual IUV and NEER (IFS)	Sample: 1980–2003. Breakpoint: No breakpoint, but countries are also sub-grouped by inflation regime, trade regime and exchange rate regimes (their Table 9).	24 developing countries (online Appendix 2).	SINGLE EQUATION Level specifications: a. level of current log (NEER) and current controls (see RHS) (no lagged dependent variable). b. a. plus differenced leads and lags in all RHS variables (no lagged dependent variable). Lag lengths not reported. (Equations estimated for each country, FMOLS and DOLS methods.)	Destination cost controls (and ‘competing price’) proxied by PPI. Exporter costs proxied by trade-weighted ‘CPI’ (they wrongly use domestic PPI instead of domestic CPI, although the IFS measures of REER and NEER use the CPI definition – see Section 5). Demand controlled for by GDP.	In their Tables 5 and 6, long-run homogeneity is not imposed and not satisfied for individual countries’ equations.	Long-run (equilibrium) ERPT for 24 countries from FMOLS and DOLS methods. Averages the above ERPT measures for groups by inflation, exchange rate and trade regimes.Test for and report no homogeneity of long-run ERPT for countries in panel regressions.
Import price aggregate and export price aggregate
Bussière and Peltonen (2008) Panel and time series Quarterly PPI, IUV and EUV (IFS, Global Insight). No separate exchange rate variable – embodied in foreign price variable, see column 5.	Sample: 1980Q1–2006Q2 maximum: some time series are short; for example, China (36 observations).Breakpoint: Stability of ERPT examined with (30-quarter) rolling regressions and breakpoint tests.	41 countries: 9 Asian; 5 L.American; 7 Central and E. European; 7 Middle Eastern and African; and 13 advanced economies(online Appendix 2) and grouped (panel): emerging/ advanced.	SINGLE EQUATION Differenced equations for IUV and EUV take the same form: a. current change in the foreign trade-weighted price (no separate exchange rate term); current change in the PPI; one lag in the dependent variable; plus current controls (see RHS).(Equations estimated for each country, OLS, and pooled)	Destination cost controls proxied by the PPI. Exporter costs proxied by trade-weighted CPI and oil price and non-oil energy price (all in foreign currency terms). No demand controls. Dummy variables for crises and hyperinflation. Note restriction: foreign prices and the exchange rate have the same coefficient – the short-run ERPT; this is not formally tested (only by simple correlations).	No long-run terms. Did not find evidence using Engle–Granger method.	For IUV and EUV models, 1-quarter (short-run) ERPT for individual countries, and averaged for all, over full sample (unbalanced panel); but foreign price variables giving ERPT elasticities are often insignificant in regressions.‘Long-run’ ERPT is quoted (defined as ERPT elasticities divided by (1- coefficient on lagged dependent variable), that is, ‘long-run’ captured by only one quarter.Bivariate cross-country regressions of short-run ERPT elasticities on macro/micro drivers.
Import price aggregate (but also CPI and PPI)
Coulibaly and Kempf (2010) Panel Quarterly CPI (IFS), PPI (IFS), IUV (IFS – 5 countries have price indices) and bilateral nominal rate against US$ (IFS, mainly).	Sample: 1989Q1- 2009Q1, but short samples in some cases; for example, start-date for IT samples: Hungary (2001:Q2), Indonesia (2005:Q3), South Africa (2001:Q1), Thailand (2000:Q2), Turkey (2006:Q1).Breakpoint: two samples with break at date of adoption of inflation targeting regime for different countries.	27 emerging countries analysed in groups: 15 inflation targeters and 12 non-targeters (online Appendix 2).	SYSTEM Differenced VAR models, 3 lags: a. (oil price, output gap, money, NER and CPI); b. 7 variables: (oil price, output gap, money, NER, import price, producer price, CPI). (Panel estimation, GMM, transforming variables as deviations from forward means to remove fixed effects.)	No destination cost controls in a. but controlled by PPI in b. Exporter costs are oil price in a. and oil price plus import price in b. Demand controlled for by the output gap. Compare impulse response functions in country subsamples, based on date of adoption of inflation targeting.	No long-run terms.	Impulse response function pictures given for ERPT to import prices, PPI and CPI before and after the adoption of targeting.Forecast variance analysis of the contribution of exchange rate shocks to inflation before and after adoption of targeting.
Ito and Sato (2008) Time series Monthly import prices (mix of prices and IUV, IFS mainly, excluding Malaysia), PPI, CPI and NEER (IFS).	Sample: Post-crisis samples: 1997/8-2006 (their Table 1).Breakpoint: No breakpoint, but analysis of the post-crisis period for Asian countries.	5 East Asian countries: Indonesia, Korea, Thailand, the Philippines, and Malaysia.	SYSTEM Differenced VAR, at most 3 lags for d.: a. (oil price, output gap, money, NEER and CPI); b. (oil price, output gap, money, NEER and PPI); c. (oil price, output gap, money, NEER and import price); d. (oil price, output gap, money, NEER, import price, PPI and CPI). Tests robustness to ordering.	No destination cost controls in a., b. or c. but controlled by PPI in d. Exporter costs are oil price in a. to d. plus import price in d. Demand controlled for by the output gap. Also control for base money or a short-term interest rate. (Prices, exchange rate and money in logs).	No long-run terms.	1-, 6-, 12-, 18- and 24- month ERPT for individual countries with accumulated impulse responses to import prices and to PPI and to CPI.Post-crisis monetary policy analysed.
Ca’Zorzi et al. (2007) Time series Quarterly import prices (mix: IUV, import deflator, prices, IFS), CPI (IFS) and NEER (IFS, BIS, OECD, JPMorgan).	Sample: 1975Q1 to 2004Q1 maximum: see samples in their Table 1, latest start-date is 1993 (Czech)). Breakpoint: No breakpoint.	12 emerging markets in Asia, Central and E. Europe, Turkey, and Latin America (online Appendix 2).	SYSTEM Differenced VAR models, up to 4 lags: a. 6 variables: (oil price, output, NEER, import price, CPI and interest rate), with interest rate in levels or differences depending on unit root tests. Tests robustness to ordering.	No destination cost controls. Exporter cost controls are oil price plus destination country import price. Demand controlled for by GDP (or industrial production index). Control also for short-term interest rate. (Prices and exchange rate in logs).	No long-run terms. Did not find evidence using Johansen co-integration tests.	4- and 8-quarter ERPT for individual countries with accumulated impulse responses to import prices and to CPI.Spearman/Pearson correlations: ERPT to consumer prices with inflation, depreciation, volatility and openness.

Notes: Abbreviations: IUV and EUV: import and export unit values; WEO: World Economic Outlook; ecm: equilibrium correction model; ERPT: exchange rate pass-through; PPP: purchasing power parity; REER and NEER: real and nominal effective exchange rates; CPI and PPI: consumer and producer price indices; Δlog (NEER): change in the log of the NEER or proportional change in NEER; RHS: right-hand-side; IT: inflation targeting; PMI: purchasing managers’ index, reflecting economic conditions; HP filter: Hodrick Prescott filter, technique for taking a long-term trend over a sample; IFS: International Financial Statistics (IMF); unbalanced panel: panel with different sample sizes for different included countries; GLS: generalised least squares; OLS: ordinary least squares; GMM: generalised method of moments; FMOLS: fully modified ordinary least squares and DOLS: dynamic ordinary least squares.

1. See online Appendix 2 for constituent countries in advanced, emerging and developing categories.

2. Where lagged dependent variables, ecm terms and interactive terms are included, considerable caution needs to be taken in calculating ERPT elasticities. If there are lagged dependent variable/s concurrent with or leading the accumulated depreciation, then extraction of the ERPT measure is no longer straightforward and an algorithm is required to extract the effects of these in the ERPT calculation (see online Appendix 1 and Choudhri and Hakura [2006, footnote 24]). If an ecm term is concurrent with or leading the accumulated depreciation, this has to be accounted for too in the ERPT calculation. Finally, if there are interaction terms, then the net effect of ERPT has to be calculated taking account of these terms, and the interacted variables need to be de-meaned so the original ERPT coefficient on is still meaningful and not affected by the different units of the interacting variables.

3. †Single equation (exchange rate assumed exogenous) or System (exchange rate is endogenous).

4. ‡The eight specialised items are alcohol (various), cigarettes, Coca-Cola, Time magazine, and Kodak film (Economist Intelligence Unit).

5. **See sections 6.2 and 6.4 for discussion of data and execution errors.

Table 2. Contrasting ERPT estimates for groups: emerging, developing and advanced countries

Study, sample and period	Advanced countries	Developing countries	Emerging countries
ERPT to trade (export and import) prices
The advanced country benchmark is: Campa and Goldberg (2005) (OECD), 1975–2003	46% (after 1 quarter)#	–	–
	64% (after 5 quarters)#
Choudhri and Hakura (2012) (18 advanced; up to16 emerging, varies in the two periods), 1979–2010	1979–2010: 60% (1 quarter)#	–	1979–2010: 54% (1 quarter)#
	1985–1997: 55% (1 quarter)#		1985–1997: 47% (1 quarter)#
	1998–2010: 49% (1 quarter)#		1998–2010: 72% (1 quarter)#
Brun-Aguerre et al. (2012) (19 advanced, 18 emerging), 1980–2009Q3 (max.)**	58%, 56%, 60% (models a,b,c after 1 quarter)‡ 74%, 100%, 47% (models a,b,c at ‘equilibrium’)‡	–	60%, 61%, 60% (models a,b,c after 1 quarter)‡ 63%, 100%, 87% (models a,b,c at ‘equilibrium’)‡
Frankel et al. (2012) (76 countries), 1990–2002, (8 specialised import products)	(41 high income)13% (after 1 year)	(DEM: for 2 low income and 15 lower-middle and 24 upper-middle income)37% (after 1 year)
Bussière and Peltonen (2008) (27 emerging), 1990Q1 to 2006Q2	Export (EUV): 22% (1 quarter)‡	–	Export (EUV): 34% (1 quarter)‡
	Import (IUV): 35% (1 quarter)‡	–	Import (IUV): 35% (1 quarter)‡
Ca’Zorzi et al. (2007) (12 emerging), 1975–2004Q1 (max.)	(For G3: US, Japan, Euro  area)	–	(For 12 markets excl. China)
	24%, 114%, 60% (4 quarters)		78% (after 4 quarters)†
	38%, 105%, 72% (8 quarters)		104% (after 8 quarters)†
Barhoumi (2006) (24 ‘developing’), 1980–2003	–	(For 9 emerging)	(For 15 developing)
		88% (equilibrium)† (DEM: for 9 emerging and 15 developing)	84% (equilibrium)†
		83% (equilibrium)‡
ERPT to CPI
The advanced country benchmark is: Gagnon and Ihrig (2004) (20 industrial countries), 1971–2000	23% (‘long run’)#	–	–
Frankel et al. (2012) (76 countries), 1990–2002	(41 high income) 2% (after 1 year)	(DEM: for 2 low income and 15 lower-middle and 24 upper-middle income) 34% (after 1 year)
Razafimahefa (2012) (SSA), 1985–2008Q2	–	20%, 41%, 42% (after 1, 4, 8 quarters)	–
Kohlscheen (2010) (8 emerging floaters), Float samples (see our Table 1)	–	–	5%, 17%, 20%, 24% (after 3, 6, 9, 12 months)#
Akofio-Sowah (2009) (SSA/LA), 1980–2005 (max.) and 2000–2005 for LA	–	(For SSA) 38% for CFA and 12% for CMA (1 quarter)	(For LA) 19% unofficially dollarised; 4% for fully dollarised (1 quarter)
Ca’Zorzi et al. (2007) (12 emerging), 1975–2004Q1 (max.)	(For G3: US, Japan, Euro area)	–	(For 12 markets incl. China)
	1%, 2%, 7% (4 quarters each)		24% (4 quarters)†
	2%, 4%, 13% (8 quarters each)		45% (8 quarters)†
Choudhri and Hakura (2006)	(after 1, 4, 20 quarters)†	(after 1, 4, 20 quarters)†	(after 1, 4, 20 quarters)†
All 71 countries, 1979–2000:	12%, 20%, 23%	15%, 24%, 27%	14 %, 24%, 27%
By low inflation sets:	8%, 14%, 16%	12%, 18%, 22%	9%, 14%, 17%
By moderate inflation sets:	24%, 39%, 42%	19%, 33%, 35%	18%, 33%, 35%
By high inflation sets:	None	30%, 46%, 52%	34%, 52%, 59%
All 71 by inflation regimes:	(after 1, 4, 20 quarters)#
	8%, 14%, 16% (low inflation regime):
	19%, 33%, 35% (moderate inflation)
	32%, 50%, 56% (high inflation)
Goldfajn and Werlang (2000)	25% (after 6 months)‡	34% (after 6 months)‡	39% (after 6 months)‡
(71 countries), 1980–98 (max.)	61% (after 12 months) ‡	51% (after 12 months)‡	91% (after 12 months)‡
	(For OECD/non-OECD, contains Turkey)	(DEM: for developing or emerging except for Denmark and Belgium)
	19% (after 12 months)‡11% (after 6 months)‡	75% (after 12 months)‡47% (after 6 months)‡

Notes: 1. The advanced economies’ benchmarks come with the caveat that biases may be present: both exclude long-run terms in a differenced reduced-form equation that assumes exchange rate exogeneity.

2. Averages: # is the quoted un-weighted average from study; ‡ is the quoted panel estimate from study; † is our calculated average (un-weighted) for this survey.

3. See online Appendix 2 for constituent countries in advanced, emerging and developing categories. For Choudhri and Hakura (2006) we allocate as follows: low income+ lower-middle (as ‘developing’); upper-middle income (as ‘emerging’), which allows near coincidence with the Emerging Markets Index list of emerging countries; and advanced countries (using the World Bank classification).

4. Brun-Aguerre et al.: their models a, b, c, (for unbalanced panel); ‘equilibrium’ for model b. is actually five-quarter ERPT (see our Table 1). Kohlsheen: plus output in the reported VAR. Frankel et al.: Equation 1 with ecm and exporters’ costs (in their Tables 8 and 10). Gagnon and Ihrig use an approximation for the ‘long run’: dividing by (1- coefficient on the one-quarter lagged dependent variable) in a differenced model – for criticism see online Appendix 1. Note that Choudhri and Hakura (2012) is a bivariate VAR without controls, and some ERPT estimates have huge standard errors.

5. **See Sections 6.2 and 6.4 for discussion of data and execution errors.

Figure 1. Decline in headline inflation for sets of countries.

Source: World Economic Outlook database (October, 2012), IMF.

Razafimahefa, I. F. 2012. Exchange rate pass-through in Sub-Saharan African economies and its determinants. Washington, DC: IMF Working Paper WP/12/141.

 Google Scholar

Kohlscheen, E. (2010). Emerging floaters: Pass-throughs and (some) new commodity currencies. Journal of International Money and Finance, 29, 1580–1595. 

 Web of Science ®Google Scholar

Reinhart, C. M., & Rogoff, K. S. (2004). The modern history of exchange rate arrangements: A reinterpretation. Quarterly Journal of Economics, 119, 1–48.

 Web of Science ®Google Scholar

Akofio-Sowah, N. A. (2009). Is there a link between exchange rate pass-through and the monetary regime: Evidence from Sub-Saharan Africa and Latin America. International Advances in Economic Research, 15, 296–309.

 Google Scholar

Choudhri, E., & Hakura, D. (2006). ERPT to domestic prices: Does the inflationary environment matter? Journal of International Money and Finance, 25, 614–639.

 Web of Science ®Google Scholar

Brun-Aguerre, R., Ana-Maria, F., & Phylaktis, K. (2012). Exchange rate pass-through into import prices revisited: What drives it? Journal of International Money and Finance, 31, 818–844.

 Web of Science ®Google Scholar

Barhoumi, K. (2006). Differences in long run ERPT into import prices in developing countries: An empirical investigation. Economic Modelling, 23, 926–951.

 Web of Science ®Google Scholar

Bussière, M., & Peltonen, T. (2008). ERPT in the global economy: The role of emerging market economies. Frankfurt: European Central Bank, ECB Working Paper No. 951.

 Google Scholar

Coulibaly, D., & Kempf, H. (2010). Does inflation targeting decrease erpt in emerging countries? Paris: Banque de France Documents de Travail 303.

 Google Scholar

Johansen, S., & Juselius, K. (1990). Maximum likelihood estimation and inference on co-integration − with applications to the demand for money. Oxford Bulletin of Economics and Statistics, 52, 169–210.

 Web of Science ®Google Scholar

Ca’Zorzi, M., Hahn, E., & Sánchez, M. (2007). ERPT in emerging markets. The IUP Journal of Monetary Economics, 4, 84–102.

 Google Scholar

Campa, J., & Goldberg, L. (2005). ERPT into import prices. Review of Economics and Statistics, 87, 379–390.

 Web of Science ®Google Scholar

Choudhri, E., & Hakura, D. (2012). The exchange rate pass-through to import and export prices: The role of nominal rigidities and currency choice. Washington, DC: IMF Working Paper WP/12/226.

 Google Scholar