1 41,2(August 2009):511–520 Journal of Agricultural and Applied Economics, 2009 Southern Agricultural Economics Association Ó Exchange Rates Impacts on Agricultural Inputs Prices using VAR Osei Yeboah, Saleem Shaik, and Albert Allen The effects of the U.S. dollar exchange rate versus the Mexican peso are evaluated for four traded nonfarm-produced inputs (fertilizer, chemicals, farm machinery, and feed) in the U.S. Unit root tests suggest that the exchange rate and the four input price ratios support the presence of unit roots with a trend model but the presence unit roots can be rejected in the first difference model. This result is consistent with a fixed price/flex price conceptual framework, with industrial prices more likely to be unresponsive to the exchange rate than farm com- modity prices. exchange rate, pass-through, law of one price, SUR, VAR Key Words: JEL Classifications: F14, F31, F36, F42, C23 to changes in exchange rates over the past de- Over the past few years, the dollar has depre- cade or so. ciated against a number of currencies. In prin- A potential decline in exchange rate pass- ciple, the dollar’s fall should help to correct the through has important implications for the U.S. U.S. trade deficit through a fall in imports, if economy. First, it has significant bearing on they are elastic. However, the dollar’s recent U.S. efforts to correct the country’s trade im- slide has produced neither a substantial fall in balance. If import prices have become much imports nor a sizable shrinking of the trade less responsive to changes in currency values, a imbalance. One possible explanation for the larger devaluation of the dollar will be needed U.S. experience of the past few years is that to narrow the imbalance. Second, pass-through the rate of exchange rate ‘‘pass-through’’—the has implications for the stability of domestic degree to which a change in the value of a prices. Low import prices are believed to con- country’s currency induces a change in the tribute to low rates of inflation—in part by price of the country’s imports and exports—has constraining domestic producers to keep their fallen relative to historical values. Indeed, prices competitive. while pass-through is almost always ‘‘incom- Though exchange rate pass-through has plete,’’ recent studies (Campa and Goldberg, long been of interest, the focus of this inter- 2005; Goldberg and Knetter, 1997) suggest that est has evolved considerably over time. After import prices in a number of industrial nations a long period of debate over the law of one may have become progressively less responsive price (LOP) and convergence across countries, beginning in the late 1980s exchange rate pass- through studies emphasized industrial organi- Osei Yeboah is associate professor of International Economics, North Carolina A&T State University, zation and the role of segmentation and price Greensboro, NC. Saleem Shai k is assistant professor of discrimination across geographically distinct h Dakota State University, Agricultural Economics, Nort product markets (Campa and Goldberg, 2006). Fargo, ND. Albert Allen is professor of Agricultural More recently pass-through studies focused on Economics, Mississippi State University, Starkville, MS.
2 512 Journal of Agricultural and Applied Economics, August 2009 and Mexico over the period 1981–2008 using an prices of traded agricultural outputs (Ardeni, vector autoregressive (VAR) model in seemingly 1989; Bradshaw and Orden, 1990; Goodwin unrelated regression (S.U.R) framework. Mex- and Schroeder, 1991; Froot, Kim, and Rogoff, ico is one of the U.S.’s major trade partners and a 1995). Adjustments of the prices of traded non- member of NAFTA. Mexico agricultural im- farm-produced agricultural inputs to the ex- ports from the U.S. grew by 228% between 1994 change rate have not received as much attention. and 2007 (post NAFTA period) while the growth Yet these purchased inputs comprise an impor- was only 25% from 1989 to 1993 (pre NAFTA tant component of agricultural production costs, membership) (FATUS, 2007). and whether their prices also respond to ex- The remainder of this paper is structured as change rate movements will affect the net im- follows: Section 2 discusses the literature re- pacts from currency revaluations. view on exchange rate pass-through; Section 3 Carter and Hamilton (1989) examined the describes a partial equilibrium framework validity of the law of one price (LOP) for traded which analyzes exchange rate effects on prices inputs used in production of wheat between the and production; Section 4 provides the theo- closely-integrated Canadian and U.S. econo- retical framework of the LOP and the specifi- mies. Over the period 1977–1986, during cations of the exchange rate pass-through which there were substantial movements in model; Section 5 discusses the development Canadian/U.S. currency values, Carter and of the VAR/SUR empirical model; Section 6 Hamilton (1989) found a contemporaneous discusses the data and estimation procedures; relationship between quarterly input prices, but Section 7 discusses the results; and Section 8 adjustments to the LOP did not occur. Also, provides conclusion of the study. while Carter, Gray, and Furtan (1990) evaluated exchange rates effects on both output and input prices, most of studies focus on output prices. Literature Review In their study, Carter, Gray, and Furtan (1990) used the LOP to examine the exchange rate This section of the paper provides information pass-through for the prices of five Canadian on several studies that provide background in- inputs—petroleum, fertilizer, pesticides, ma- formation on the impact of exchange rates on chinery, and fat steers—and three Canadian prices. The articles reviewed in this study serve outputs—wheat, canola, and feeder steers— as a selective set of articles by the authors. using quarterly data over the period 1975– Abeysinghe and Yeok (1998) used an 1988. Carter, Gray, and Furtan (1990) found econometric model to estimate the effects of that the exchange rate had significant pass- import content on exports and the dynamic through effects on some of the input prices as effects of productivity improvements on the well as the output prices, although differences competitiveness of Singapore’s exports. Results occurred in the timing and extent of this reveal that, in general, the higher the imported pass-through. More recently, Carlson, Deal, input content, the less impact of exchange rate McEwan, and Deen (1999) have provided a changes on exports. At one extreme, exchange descriptive analysis of the relationships be- rate changes had no effect on re-exports, while tween herbicide prices in Canada and the U.S. at the other extreme, service exports, being using cross-sectional annual data over the period relatively less intensive than imported inputs, 1993–1999. Carlson et al. (1999) concluded that were most affected by currency exchanges. The restrictions on the movement of pesticides authors further found that productivity gains across the border are one factor creating price were not sufficiently large enough to contribute differentials for similar products. significantly to enhance export price competi- This study develops a system of empirical tiveness. This result suggested that domestic models that capture the short-run dynamics of value-added was not as significant as imported exchange rate and the LOP effects on four traded input content in influencing export prices. nonfarm produced inputs (chemicals, farm ma- Byrne, Darby, and MacDonald (2008) chinery, feed and fertilizers) between the U.S. measured the impact of exchange rate volatility
3 513 Yeboah, Shaik, and Allen: Exchange Rate Impact on Agricultural Inputs Prices across both sectoral activity and prices in the on the volume of bilateral U.S. trade (both euro area. Overall, the sector results suggest that exports and imports) using sectoral data. The within industry (excluding construction), the authors used bilateral imports from and exports main industrial groupings (MIGs), capital and to the U.S. from a sample of six European intermediate goods, account for almost all of the countries. In this analysis, the authors used impact on production (around 90%), while disaggregate price data as the trade deflator, among the main subsectors the whole impact rather than the U.S. consumer price index comes via the manufacturing sector. On the (CPI), and they constructed new disaggregate price side, the most important contributor to the sectors to examine the importance of exchange effect on producer prices in industry among rate uncertainty. Results reveal that pooling MIGs is the energy sector, accounting for more all industries together provides evidence of a than 50% of the overall effect, while among negative effect on trade from exchange rate subsectors the largest contribution may be volatility. However, when the authors used an ascribed to producer prices in manufacturing; econometric model, they found evidence that however, in contrast to the effects on activity, this effect may be different across industries. In the electricity, gas, and water supply sector addition, the authors found that output and contributes significantly. relative price coefficients are different on a Parsley and Popper (2006) reexamined de- disaggregated basis. Moreover, the effect of compositions of the real exchange rate that exchange rate uncertainty is negative and sig- apportioned its movements into a part that nificant for differentiated goods, and insignifi- reflected international deviations from the law cant for homogeneous goods. of one price and a part that reflected the relative Campa and Goldberg (2006) found that prices of traded and nontraded goods within border prices of traded goods are highly sen- countries. Using a partial equilibrium model sitive to exchange rates; however, they found with Japanese and U.S. data, the authors showed that the CPI and the retail prices of goods that that in such decompositions the traded/non- make up the CPI are more stable. The authors traded distinction was irrelevant at the con- decomposed the sources of that price stability sumer level. Also, the authors, motivated by a for 21 OECD (Organization for Economic model of trade in intermediate products, used Cooperation and Development) countries, fo- implied import weights and found that relative cusing on the important role of distribution traded/nontraded price changes accounted for margins and imported inputs in transmitting much of the real exchange’s rate variation. exchange rate fluctuations into consumption Parsley and Wei (2003) studied the move- prices. The authors found that distribution ment of real exchange rates based on prices of costs, relevant to consumer price pass-through Big Macs. The authors matched these prices to calculations, were on average 32–50% of the the prices of individual ingredients (ground total costs of goods across OECD countries. beef, bread, lettuce, labor cost, rent and other The authors also found that imported input use items) in 34 countries during 1990–2002. Re- is larger in tradable goods industries than in sults showed that the nontraded component of nontradables production, and varied widely Big Mac prices was substantial, ranging be- across countries. tween 55–64%. The authors also studied the Hahn (2007) investigated the impact of ex- persistence of the real exchange rate in a set- change rate shocks on sectoral activity and ting free of possible biases induced by non- prices in the euro area. Using a VAR frame- comparability of consumption baskets across work, the author provides evidence on the countries, product aggregation bias, and time magnitude and speed of the impact of exchange aggregation bias. The authors found that the rate shocks on activity in all main euro area speed of convergence for the Big Mac real ex- sectors and on the activity and producer prices in change rates was slower than the speed for its a large set of subsectors of industry. The results tradable inputs, but faster than for its non- from this analysis suggest a high degree of tradable inputs. Finally the authors showed that heterogeneity in the exchange rate sensitivity
4 514 Journal of Agricultural and Applied Economics, August 2009 world prices. The currency depreciation may Engel’s result that deviations from the law of one then increase traded input prices— P price are all that matters does not hold generally to 1 (Engel, 1999). P —and thus the cost of production, in the 2 longer run. If all of the inputs are traded as in this study and there is eventually a complete Partial Equilibrium Analysis exchange rate pass-through to their costs, then 1 output supplied would remain unchanged at Q Devaluation in the exchange rates means an after full adjustment to the depreciation. In the increase in the nominal and real prices in the case that not all inputs are traded, or that ex- tradable sector. When the domestic currency change rate pass-through effects on input prices depreciates it increases the traded commodity are incomplete, output supplied would be de- price, but its impact on supply also depends 2 by factors in- Q 1and Q termined between on input price changes. If a fixed price/flex cluding the elasticity of the supply function, the price model is assumed (Saghaian, Reed, and proportion of traded inputs in production, and Marchant, 2002), then output prices respond output responses to changes in the input prices. contemporaneously to exchange rate move- ments while traded input prices are unrespon- Model Specification of Exchange Rate sive in the short run. But inputs may also be Pass-Through and LOP traded if the home country is assumed to import at least some inputs from the foreign country. The law of one price (LOP) states that in the However, in this study, Mexico (the foreign absence of transportation and other transaction country) imports agriculture inputs from the costs, competitive markets will equalize the U.S., such as machinery and agriculture chem- prices of an identical good in two countries icals. When the domestic currency depreciates, when the prices are expressed in the same the prices of goods imported into that country currency. In mathematical notation, the law of are typically expected to rise. one price can be expressed as follows: The underlying partial equilibrium frame- f d work to examine the effects of exchange rate EP 5 (1) P t , i t , i changes on small specific industries using a f simple model of the firm is developed and also d P P where and are the domestic and the t t presented graphically in Figure 1. The primary corresponding foreign currency price respec- assumptions include the exogenous nature of the i for the time period t and tively of a commodity exchanges rates and that the countries are large is the exchange rate defined as the domestic- E nations, i.e., both countries’ trade has impacts on currency price of foreign currency. Given transportation and storage costs and the imperfect competitive world market, the absolute version of the LOP as expressed in Equation (1) is very unlikely to hold. However, the following relative version of the law of one price may hold: f d P 5 (2) EP a , t i , t i a indicates the deviation from the law of where one price, and is constant over time. Equation (2) can be rewritten as: d P i t , df a 5 [ (3) P E , t i f P i , t Figure 1. Effects of Exchange Rate Depreci- with the advent of time-series analysis, VAR ation and vector error correction (VEC) processes
5 515 Yeboah, Shaik, and Allen: Exchange Rate Impact on Agricultural Inputs Prices r s X X have gained popularity due to their flexibility df df a E 5 1 P b (5) P j t i i j t , , t and ability to estimate relationships of the 1 j 5 0 5 j variables of interest for stationary and non- hi 9 stationary with cointegration correction, re- df df df df , , P P 5 P , represent vec- P P where , t , t , t , t 4 1 3 2 spectively. Theories examining the short and long run relationships between exchange tors of endogenous variables, and E is the exog- are lags in exogenous and enous variable; s and r rates and domestic prices have been examined in a dynamic framework (Carter, Gray, and autoregressive components respectively. Furtan, 1990; Chambers and Just, 1981). Ear- Next, this study develops a system of equa- tions estimation model that captures the short- lier studies have examined independently the relation between each input price and exchange run dynamics of U.S. vs. Mexican exchange rate rate using VAR or VEC process. effects on U.S. input prices using an SUR model. The SUR/VAR representation of input prices and Equation (3) is used to examine the impor- tance of exchange rate on four input prices, and exchanges rate for Equation (5) is: d d d s s r X X X P P P fert feed feed t , , j j t , t ln a 1 b 1 ln E 5 ln b ln 1 j j , j , 11 12 j t f f f P P P j 0 5 j 1 1 5 j 5 , j j t , t , t fert feed feed d d s s X X P P chem mach , t , j t j b 1 ln ln 1 e b 1 1 , 13 j j 14 , f f P P 0 j 5 5 j 1 mach chem t , j , t j d d d r s s X X X P P P , , t j j t , t fert fert feed ln ln a 1 ln 5 b 1 E ln b 2 j 21 , 22 j j , j t f f f P P P j 5 1 0 5 j 1 5 j fert fert feed t t , , , j t j d d s s X X P P mach chem j , t t , j ln ln b 1 e 1 b 1 2 , , j 23 24 j f f P P 0 j 1 5 j 5 chem mach , t , j j t (6) d d d s s r X X X P P P t , j t , j , t chem fert feed ln ln 1 E ln b ln 1 b a 5 3 j 32 j j , , 31 t j f f f P P P j 5 1 j j 1 5 0 5 fert chem feed j t t , t j , , d d s s X X P P chem mach t j , , t j ln ln b b e 1 1 1 3 j , 34 j , 33 f f P P 5 1 j 5 0 j chem mach j , j t , t d d d s s r X X X P P P fert feed mach t j , , t t j , ln ln a b 5 1 1 ln b E ln j 4 41 , j 42 j , t j f f f P P P 0 j 1 5 j 5 5 j 1 fert feed mach t , j t , , t j d d s s X X P P mach chem , j , t t j ln ln b b 1 1 e 1 4 44 j , 43 j , f f P P 5 j 0 5 j 1 chem mach , t j j t , hence can be rewritten as: is years; a where t and b are estimated param- eters associated with exchange rate and lagged s e endogenous variables; and e and e , are e , X 4 1 2 3 df E a 5 P (4) j t , t i errors for each of the four input price equations. 0 5 j Parameter estimates from Equation (6) would still allow us to recover the short-run relationships between exchange rate and the are where E is the exogenous variable and s four endogenous price variables. lags in exogenous component. As we are interested in examining the short Data and Estimation Procedures analysis, the VAR model of Equation (4) with contemporaneous and lagged exogenous and Agri- The input price series are derived from lagged endogenous variables can be repre- published by the National cultural Prices sented as:
6 516 Journal of Agricultural and Applied Economics, August 2009 Summary Statistics of the Variables Used in the Analysis Table 1. Maximum N Mean Std Dev Median Minimum Variable 323 Chemicals 105 32.6 74.4 2.53 0.74 2.23 327 Machinery 105 31.2 75.5 0.69 1283 105 99.8 265.1 2.21 0.76 Feed 72.2 Fertilizer 105 188.3 2.80 0.95 890 0.058 0.018 0.188 U.S. MexicoEX 105 0.093 0.095 Results Agricultural Statistics Service (NASS) of the USDA. The U.S./Mexican exchange rate is compiled by the Economic Research Service Table 1 presents the summary statistics of input price indices and the exchange rate of the U.S. (ERS) of the USDA. Monthly data are con- dollar vs. the Mexican peso. The mean index verted to quarterly averages for consistency in the analysis, since the input price series are for chemicals for the 105 quarters is 32.6 with a minimum of 0.74 and maximum of 323 while only available on a quarterly basis. Data on farm machinery has 31.2 as mean index and exchange rates were obtained from the For- eign Agricultural Trade of the United States 0.69 and 327 minimum and maximum. Thus chemicals and machinery have almost the same (FATUS) database on the USDA’s ERS web- range as indicated by the standard deviations of site. The exchange rate data are measured as 74.4 and 75.5. Feed and fertilizer indices are the U.S. dollar per the Mexican Peso, which completely different from each other and from means that an increase indicates a deprecia- both chemicals and machinery. The means of tion of the U.S. dollar, and a decrease means these indices are 99.8 and 72.2 with minimums depreciation. The parameter(s) of Equation (7) of 0.76 and 0.95, respectively, whereas the is estimated in dynamic model accounting maximums are as large as 1283 and 890. The for the system of equations. This consists of mean exchange rate of the dollar to the peso for first choosing the optimum lag using Akaike the study period is about 9 cents with a mini- Information Criteria (AIC) by estimating an mum and maximum of about 19 cents to the unrestricted model with one lag of each en- peso. dogenous variable. Based on the AIC model Results of the unit root tests are presented selection, the specification included one lag in Table 2. The results in Table 2 indicated for all the endogenous variables. Due to the that all of the four input price ratios—feed, use of quarterly data, we include four lags of fertilizer, machinery and chemicals—support exogenous exchange rate variable for each the presence of unit roots with a trend of the four input price equation. The dynamic model. The results of the first difference in- model with one lagged endogenous variable dicate that the presence of unit roots can be of all four input prices and four lagged exoge- rejected. nous exchange rate in each equation is esti- To account for unit roots, the SUR/VAR mated using with iterative SUR system of model defined in Equation (6) is estimated equations. using the first difference of exogenous and For a complete exchange rate pass-through endogenous variables. The SUR/VAR model and adherence to the LOP, we hypothesized was estimated with four lags for exogenous the sum of the coefficients of the contem- components and just one lag for the autore- poraneous and that of the lags sum up to gressive. Use of higher lag for the autore- one, whereas a sum equal to zero represents gressive was avoided as the model did not the null hypothesis which implies no ex- converge due to high colinearity. Equation (6) change rate pass-through and invalidity of the can be rewritten roots as: LOP.
7 517 Yeboah, Shaik, and Allen: Exchange Rate Impact on Agricultural Inputs Prices df df D 5 E P D a b 1 1 a D E D E 1 a 1 a D P 1 12 j 3 11 11 1 t 2 t t feed feed , , t 1 t df df df e 1 1 1 b D P b P D D b 1 P 1 14 13 12 chem mach fert t 1 t , t 1 , 1 , df df 1 P D 5 a b 1 a 1 D E D 1 a E D E D a P 23 22 21 2 21 1 2 t t t feed fert , t , 1 t df df df e P D b 1 P D b 1 1 P D b 1 2 22 24 23 chem fert mach 1 , t 1 , , t 1 t (7) df df D P D E D a 1 b 1 1 a 1 a 5 E D a E P D 31 32 33 3 31 t 2 t t 1 chem feed t , t 1 , df df df P D b 1 1 P D b 1 P D b 1 e 3 32 34 33 mach chem fert 1 1 t t , , 1 t , df df E D 1 1 1 a P D 5 a a D D P E 1 a E D b 43 42 41 4 41 t t 1 t 2 mach feed t 1 , t , df df df D 1 b 1 1 b P D P b D e P 1 4 44 42 43 fert chem mach , 1 t , , t 1 1 t df f d d rate, especially in the feed and fertilizer equa- / P / ln ln D 5 Þ P ð ð P P where, , t , t 1 feed feed feed feed , t , t f tions, where none of the elasticities of the lags P Þ , is the first difference of the ratio , 1 t feed are significant. Also, none of the contempora- of the feed price between the domestic and neous point estimates are statistically signifi- foreign country. Similarly the first difference of cant. Most importantly, we fail to reject the null the exchange rate ratio between the domestic hypothesis that the coefficients of the contem- E and foreign country is defined as D 5 t poraneous and those of the lags sum up to zero E ln E . Even though it is possible to ln t 1 t which implies no exchange rate pass-through test for the presence of autocorrelation and and invalidity of the LOP. Exchange rate pass- heteroskedasticity for each equation, it is through is limited for all the inputs—fertilizer, more appropriate to test for autocorrelation feed, chemical, and farm machinery—even and heteroskedasticity in an SUR/VAR frame- after four quarters. The sums of coefficients are work to account for possible error correlation 0.70 for feed, 0.08 for fertilizers, and 0.45 for across equations (Breusch and Pagan, 1979; machinery, while that of chemicals is 0.00. White, 1980; and Godfrey, 1978). Results did These results are not exceptional as most not indicate the presence of autocorrelation or studies (Carter, Gray, and Furtan, 1990; Xu heteroskedasticity. and Orden, 2002) find input prices to be sticky. The estimated contemporaneous, one-lag up Xu and Orden (2002) finds the farm pass- to four-lag coefficients of the VAR/SUR model through effect on farm machinery to be only are presented with t -statistics Table 3. In all, the 0.37, even after two years, suggesting that price input prices show a less response to exchange Dickey-Fuller Unit Root Tests of the Variables Table 2. Pr < Rho Variable Pr < Tau Type Rho Tau Levels 2 2.26 0.9615 2 1.76 0.7175 Chemicals Trend 2 3.71 2 2.11 0.5355 Machinery Trend 0.8998 2.4 0.8161 2 1.52 2 Feed Trend 0.9572 2 0.9792 2 1.28 0.8861 Trend Fertilizer 1.53 2 19.78 0.059 2 3.2 0.0907 MexicoEX1 Trend First Difference Trend 54.2 0.0003 Chemicals 5.12 0.0003 2 2 2 0.0003 2 4.79 0.0009 Trend Machinery 44.76 2 85.27 0.0003 2 6.43 <0.0001 Feed Trend Trend 2 0.0003 2 6.12 <0.0001 Fertilizer 77.4 Trend 105.47 0.0001 2 8.91 <0.0001 MexicoEX1 2
8 518 Journal of Agricultural and Applied Economics, August 2009 Exchange Rate Pass-Through for U.S. Agricultural Inputs Table 3. t Value Estimate t Value Estimate D Variable D feed fertilizer 0.07 Intercept 2 0.0005 2 0.06 0.0006 0.84 0.48 D USMexicoEX(t) 0.0946 0.0490 USMexicoEX(t-1) 0.1814 1.52 0.1185 1.11 D D 1.22 USMexicoEX(t-2) 0.2879 2 0.0505 2 0.24 2 2 D USMexicoEX(t-3) 0.1432 1.04 0.28 0.0342 USMexicoEX(t-4) D 0.52 0.0481 0.52 0.0534 D 2 0.1411 2 0.99 0.1110 0.87 chemicals(t-1) D machinery(t-1) 0.0059 0.02 2 0.1351 2 0.59 D feed(t-1) 2 0.3744 1.81 0.1643 3.7 2 0.08 0.0106 D 4.43 2 0.5245 2 fertilizer(t-1) chemicals D machinery D Intercept 2 0.0005 2 0.08 2 0.0011 2 0.23 D USMexicoEX(t) 0.0258 0.34 0.1485 2.35 D USMexicoEX(t-1) 0.2216 2.78 0.1386 2.07 D 2 0.3378 2 2.14 0.0910 0.69 USMexicoEX(t-2) USMexicoEX(t-3) 0.47 D 0.0715 0.93 0.0432 D 0.1168 1.7 2 0.0002 0 USMexicoEX(t-4) D chemicals(t-1) 2 0.5523 2 5.81 0.0015 0.02 D 2 0.1016 2 0.59 2 0.4910 2 3.42 machinery(t-1) feed(t-1) D 2.55 0.1042 1.54 0.1441 D fertilizer(t-1) 0.0903 1.02 2 0.0277 2 0.37 Conclusion adjustment to exchange rate movements re- mains incomplete. This paper investigates the effects of the U.S. Only four lagged regression coefficients are dollar exchange rate versus the Mexican peso a significant in the VAR/SUR model ( and 1 t 2 on the prices of four traded nonfarm-produced a a and ) for both chemicals and ( a ) farm 1 t 2 t 2 2 t inputs (fertilizer, chemicals, farm machinery, machinery. A 1% depreciation of the dollar and feed) in the U.S. Unit root tests suggest that today raises the prices of chemicals by 0.22% the exchange rate and the four input price in three months while the prices fall by 0.34 in ratios—feed, fertilizer, machinery and chem- the sixth month. This result can be explained by icals—support the presence of unit roots with a ‘‘J-curve effects.’’ Due to the lagged adjust- trend model but the presence of unit roots can ments in trade volume on prices changes, a be rejected in the first difference model. To depreciation will reduce export values and in- account for unit roots and the system of four crease import values which will trigger infla- inputs, a VAR model in SUR framework was tionary conditions before prices fall to improve developed to identify the importance of ex- trade balance. For farm machinery, a 1% de- change rates on agricultural inputs. Further, preciation of the dollar contemporaneously in- the autocorrelation and heteroskedasticity for creases the price by 0.15% and 0.14% in three the four system of inputs was tested in SUR months. For feed, although the estimated pass- framework along with a VAR model. through increases over time, the evidence is not The empirical results confirm that short-run strong enough to reject either the null hypoth- adjustments to the LOP do not occur even after esis of zero exchange rate effect or the LOP. For five quarters for all of the agricultural input chemicals, LOP and zero pass-through are prices. Therefore, the LOP is refuted for all strongly rejected.
9 519 Yeboah, Shaik, and Allen: Exchange Rate Impact on Agricultural Inputs Prices 72,3(August 1990): Agricultural Economics four inputs. This result is consistent with a 738–43. fixed price/flex price conceptual framework Carter, C.A., and N.A. Hamilton. ‘‘Wheat Inputs with industrial prices more likely to be unre- Agribusiness and the Law of One Price.’’ sponsive to the exchange rate than farm com- 5(1989):489–96. modity prices. Carter, A., R.S. Collin, and W.H. Gray. ‘‘Furtan Future research looks forward to extend- American Journal of Agricultural Source.’’ ing the analysis to specific inputs that are 72,3(1990):738–43. Economics traded extensively and insignificantly; exam- Chambers, R.G., and R.E. Just. ‘‘Effects of Ex- ining the robustness of the results under the change Rate Changes on U.S. Agriculture: A presence and absence of cross input equation Dynamic Analysis.’’ American Journal of Ag- ricultural Economics 63(1981):32–46. correlation; and finally, extending the analysis Engel, C. ‘‘Accounting for U.S. Real Exchange under VEC framework to account for the Journal of Political Economy Rate Changes.’’ cointegration. 107(1999):507–38. Euromonitor International . 30th ed. International Marketing Data and Statistics, 2006. Foreign Agricultural Trade of United States References (FATUS). (2006) USDA Economic Research Service (ERS). Internet site: www.ers.usda. Abeysinghe, T., and T.L. Yeok. ‘‘Exchange Rate gov/. Appreciation and Export Competitiveness: The Foreign Agricultural Trade of U.S. (FATUS). Case of Singapore.’’ Applied Economics (2007) USDA Economic Research Service 30(1998):51–55. (ERS). Internet site: www.ers.usda.gov. Ardeni, P.G. ‘‘Does the Law of One Price Really Froot, K.A., M. Kim, and K. Rogoff. ‘‘The Law of American Hold for Commodity Prices?’’ One Price Over 700 Years.’’ National Bureau of 71(1989): Journal of Agricultural Economics Economic Research Working Paper 5132, May 661–69. 1995. Bradshaw, G.W., and D. Orden. ‘‘Granger Cau- Goldberg, P.K., and M.M. Knetter. ‘‘Goods Prices sality from the Exchange Rate to Agricultural and Exchange Rates: What Have We Learned?’’ Prices and Export Sales.’’ Western Journal of 35,3(1997): Journal of Economic Literature Agricultural Economics 15(1990):100–10. 1243–72. Breusch, T.S., and A.R. Pagan. ‘‘A Simple Test for Godfrey, L.G. ‘‘Testing for Higher Order Serial Heteroscedasticity and Random Coefficient Correlation in Regression Equations when Econometrica Variation.’’ 47,5(1979):1287–94. the Regressors Include Lagged Dependent Byrne, J.P., J. Darby, and R. MacDonald. ‘‘US 46(1978a):1303–10. Variables.’’ Econometrica Trade and Exchange Rate Volatility: A Real ———. ‘‘Testing Against General Autoregres- Sectoral Bilateral Analysis.’’ Journal of Mac- sive and Moving Average Error Models roeconomics 30(2008):238–59. when the Regressors Include Lagged Depen- Campa, J.M., and L.S. Goldberg. ‘‘Distribution Econometrica 46(1978b): dent Variables.’’ Margins, Imported Inputs, and Sensitivity of 1293–301. the CPI to Exchange Rates.’’ Federal Reserve Goodwin, B.K., and T.C. Schroeder. ‘‘Price Dy- Bank of New York Staff Reports, Staff Report namics in International Wheat Markets.’’ Ca- No. 247, April 2006. nadian Journal of Agricultural Economics ———. ‘‘Exchange Rate Pass-Through into Im- 39(1991):237–54. port Prices.’’ The Review of Economics and Hahn, E. ‘‘The Impact of Exchange Rate Statistics 87,4(2005):679–90. Shocks on Sectoral Activity and Prices in the Carlson, G., J. Deal, K. McEwan, and B. Deen. Euro Area.’’ European Central Bank Euro- ‘‘Pesticide Price Differentials between Canada system Working Paper Series No. 796, August and the U.S.’’ Prepared for Economic Research 2007. Service (ERS) of the United States Department Parsley, D., and H. Popper. ‘‘Understanding Real of Agriculture (USDA) and Agriculture and Exchange Rate Movements with Trade in In- Agri-Food Canada, Fall 1999. termediate Products.’’ (2006). Internet site: http:// Carter, C.A., R.S. Gray, and W.H. Furtan. ‘‘Ex- papers.ssrn.com/sol3/papers.cfm?abstract_id 5 change Rate Effects on Inputs and Outputs in American Journal of Canadian Agriculture.’’ 904323 (Accessed November 16, 2008).
10 520 Journal of Agricultural and Applied Economics, August 2009 U.S. Department of Agriculture (USDA), Eco- Parsley, D., and S. Wei. ‘‘The Micro-Foundations Agricultural nomic Research Service (ERS). of Big Mac Real Exchange Rates.’’ Hungarian . Selected issues. Outlook Academy of Sciences Institute of Economics White, H. ‘‘A Heteroskedasticity-Consistent Co- Discussion Papers, New Series MT-DP 2003/6, variance Matrix Estimator and a Direct Test for Budapest, Hungary, June 2003. Heteroskedasticity.’’ 48,4(1980): Econometrica Saghaian, S.H., M.R. Reed, and M.A. Marchant. 817–38. ‘‘Monetary Impacts and Overshooting of Xu,M.,andD.Orden.‘‘ExchangeRate Agricultural Prices in an Open Economy.’’ Effects on Canadian/U.S. Agriculture.’’ Pa- American Journal of Agricultural Economics per presented at the annual meeting of the 84,1(February 2002):90–103. American Agricultural Economics Associa- U.S. Department of Agriculture (USDA), Na- tion, Long Beach, California, July 28–31, tional Agricultural Statistics Service (NASS). 2002. Agricultural Prices . Selected issues.