Local agricultural markets and external shocks – the case of Poland

According to Eurostat data, in 2021 Poland ranked fourth in the EU in terms of grain export volume (after France, Romania, and Germany). In the structure of Polish grain exports in 2015–2021, wheat and corn were the most important, accounting for 53.4% and 20.5% of its volume, respectively. It is worth noting the increasing share of corn in Polish exports of grains, which is a result of the above-mentioned growing domestic production.

Due to the geographical conditions of Poland, the domestic production of oilseeds is dominated by the production of rapeseed. In the years 2015–2021, its average annual production amounted to ≈2.7 million tons, which was the third highest in the EU (after France and Germany). The average annual consumption of rapeseed in Poland in 2015–2021 amounted to 2.7 million tons, and in the analyzed period it was in an upward trend. Thus, although at the beginning of the analyzed period Poland recorded a surplus in rapeseed trade, in recent years it has become its net importer. It should be noted that the domestic demand for high-protein crops is not limited only to rapeseed, and due to the growing domestic animal production, Poland is heavily dependent on imports of oilseed meals (in particular, soybean meal) and vegetable oils.

The concept of price transmission has its roots in the Law of One Price. According to this concept, the price difference between the same goods in two different markets is at most equal to the trade costs between these markets [Baffes, 1991; Mundlak and Larson, 1992; Fackler and Goodwin, 2001]. Economists have noticed that price transmission depends on many factors. We see this in the extensive literature on price transmission. Six factors can be distinguished as follows [Olipra, 2020]: (1)

Product uniformity [Minot, 2010];

To sum up, price transmission research focuses on several different areas, the main ones being as follows [Olipra, 2020]:

The impact of market deregulation and trade liberalization on the transmission of prices on agricultural markets;

The subject of the impact of exchange rates on the prices of agricultural products is extremely rich. For example, according to Schuh [1974, 1984], the exchange rate plays a key role in all aspects of agriculture. Bessler and Babula [1987] assessed the effects of changes in exchange rates on the prices and sale of wheat, including transport prices. Bradshaw and Orden [1990], on the other hand, using Granger causality, tested the effect of exchange rates on agricultural prices and exports. Most studies confirm that exchange rates are an important factor that determine the prices, supply, and demand in the agricultural market [Paarlberg et al., 1994]. While the very impact of exchange rates on the processes of setting prices for agricultural products and the value of trade is indisputable, there is still no consensus about how important their role is. It is pointed out that their impact on individual commodity groups differs [Tweeten, 1992], and that the sensitivity of grains to exchange rate changes depends on whether more of them are exported than consumed domestically [Rausser et al., 1986]. The attention of economists was also drawn to the issue of the impact of monetary shocks (a sharp increase in the money supply) on the prices of agricultural products. For example, Orden and Fackler [1989] showed that a shock change in the money supply also contributes to the volatility of agricultural prices.

In addition, it is also an open question to determine the macroeconomic variables that may affect price processes in the agricultural market. Empirical studies show different results in this case. A broad discussion of this issue can be found in a literature review on the relationship between exchange rates and prices of agricultural products, by Kristinek and Anderson [2002]. Their study discussed in detail the issue of the role of the exchange rate in agriculture, emphasizing that most studies confirm the relationship between exchange rate changes and prices of agricultural products and their exports. An interesting study was conducted by Johnson et al. [1977], who compared the effect of the exchange rate with the effect of foreign trade policy on US wheat prices. They found that foreign trade policy, designed to protect consumers from rising prices, had a greater impact on domestic wheat prices than US monetary policy. They determined that the devaluation of the dollar had a positive effect on domestic wheat prices through increased export demand and thus lower domestic supply.

The results of research conducted by Rembeza and Seremak-Bulge [2009] draw attention to the analysis of price transmission between domestic and foreign markets. Their study for the years 1996–2008 shows that domestic and foreign grains prices are co-integrated. Foreign prices strongly influence corn prices in Poland, but there was practically no impact of Polish grains prices on foreign ones.

Various types of econometric models are used in studies of the situation on the grain market. Often, these are models of extrapolation of the trend function, which are the simplest way to forecast phenomena characterized by a trend. In the case of time series of prices of agricultural products, medium-term deviations in the form of quasi-cyclic fluctuations can be observed. Therefore, models extended with an autoregressive component are used. In addition, adaptive models are also used, where the parameters of the models adapt to changes over time. This group of models includes exponential smoothing models, the Holt model with a fading trend. Models with explanatory variables, VAR and VECM vector autoregressive models, and partial equilibrium models of the agricultural sector are also used for forecasts.¹

They include classic time series decomposition,² regARIMA methods, simple Holt model, or Holt model with a smoothing factor.³ It should also be noted that the frequently highlighted problems relating to the time series of prices of agricultural products are their lack of stationarity and the occurrence of seasonality. However, the methods do not take into account long-term relationships and the assessment of the impact of these relationships in response to a given shock.

To sum up, in the case of Poland, more detailed analysis is required to quantify the factors that influence the domestic prices of agricultural products. Moreover, it should be checked whether in the event of external shocks, factors such as a favorable supply situation in the country may isolate the local market.

The analysis of domestic prices of grains and oilseeds with the use of error correction models was carried out using a two-stage procedure proposed by Engel and Granger [1987].⁴ The parameters of the error correction model were estimated in the following form: M.1Δyi,t=β0+θ(yi,t−1−α1xj,t−1−α2fxt−1−α0)+β1Δyi,t−1+β2Δxj,t−1+β3Δfxt−1+ϵt where:

y_i,t – natural logarithm of commodity purchase prices i in period t;

In a properly defined long-term relationship, parameters α₁,α₂ should be positive, which is identical to the assumption that in the long run, the prices of futures contracts for a given agricultural commodity, the EUR/PLN, and domestic prices of a given agricultural commodity remain in a constant relationship. Its occurrence can be interpreted as a stable, long-term relationship to which the variables return when they deviate from equilibrium. This means that in the long term, domestic prices of the analyzed grains and oilseeds are determined by the prices of futures contracts and the level of the EUR/PLN. Similarly, it can be expected that the short-term parameters β₁, β₂ should be positive, which means that an increase in futures prices or depreciation of PLN leads to an increase in domestic grain prices. In the error correction model, the parameter θ should be negative, which ensures that deviations from the long-term relationship are gradually corrected. The closer the parameter θ is to -1, the faster the rate of return to this relation is, while the closer it is to 0, the slower the rate of return to equilibrium.

The obtained estimation results of the error correction models (Table 2) indicate as follows:

Wheat: There is a long-term co-integration relationship between domestic wheat prices and wheat futures prices and the level of EUR/PLN. Deviations of domestic wheat prices from the long-term relation are systematically corrected. The point value of the parameter θ = –0.245 means that 24.5% of these deviations are corrected within 1 month. Short-term parameters indicate that a 10% increase in wheat futures prices increased domestic wheat prices by ≈2.1%. Among the three analyzed grains, the short-term impact of the exchange rate on domestic prices was found only in the case of wheat.

In this part of the study, the parameters of the reaction function of domestic prices of grains and oilseeds were estimated to changes in the prices of futures contracts, and then the parameters of the reaction function were also analyzed, distinguishing between the level and volatility of EUR/PLN.

In the analyzed models of the reaction function of the domestic prices of wheat, maize, and rapeseed in PLN, the main determinants were the prices of futures contracts for these agricultural commodities on the European market (in EUR), the exchange rate, the sea freight price index, and the volume of domestic harvest of grains and oilseeds. The base specification of the model is as follows: where: M.2 Δyi,h=yi,t+h−yi,t−1=α0+α1,hΔxi,tF+α2 fx t+α3 freight t+α4 crop t++α5 month +εt where:

y_i,t+h – natural logarithm of domestic commodity purchase prices i in PLN, in period t + h, where h = 0,1,…,6;

Among the control variables in the M.2 model, the first lags of changes in domestic prices of grains and oilseeds, futures prices, EUR/PLN, and changes in the freight price index were taken into account. In addition, the model specification includes changes in the prices of futures contracts for these agricultural commodities, the level of exchange rate, and changes in the sea freight price index for each analyzed horizon h. The model estimates obtained in this way are not biased by changes in these parameters in the analyzed horizon. For example, for h = 2, the model specification for wheat includes changes in wheat futures prices, EUR/PLN, and changes in the freight price index in the period t + 1 and t + 2. In the model M.2, the dependent variables were defined as the difference in the logarithms of the prices of the relevant agricultural commodities, determined separately for each time horizon h = 0, 1, …, 6. As a result, the parameter α_1,h should be interpreted as the cumulative change in prices of commodity i in the horizon h for the change in prices of futures contracts for commodity i on the European market in the period t = 0. Since we wanted to verify the impact of the exchange rate and futures prices on the development of domestic prices of agricultural commodities, the estimation results are supplemented with charts illustrating the impact of the hypothetical depreciation of PLN against EUR by 1% and an increase in futures prices by 1% on the domestic prices of wheat, corn, and rapeseed within 6 months.

The results of the estimation of the reaction function (Table 3) indicate as follows: (a)

Wheat: Changes in wheat futures prices are very quickly reflected in changes in domestic wheat prices: a 1% change in wheat futures prices causes an immediate (i.e., at t = 0) increase in domestic wheat prices by about 0.5%, and from the following month (i.e., from period t = 1, …, 6) also by ≈1%. The depreciation in EUR/PLN raises domestic wheat prices, but these effects take place with a 1-month lag. The increase in domestic wheat prices does not prove to be driven by changes in the sea freight prices. The cumulative change in domestic wheat prices in response to the depreciation of the PLN against the EUR is statistically different from 0 from the first month of the projection and the response to futures prices turns out to be significant immediately (i.e., from month 0).

In this part of the study, it was verified whether the level and volatility of the exchange rate affect the parameters of the response function of domestic prices of grains and oilseeds to changes in futures prices. The level and volatility of the exchange rate refer here to the average values of these measures observed in the analyzed period, that is, January 2015–October 2023. The basis for inferring the impact of the situation on the currency market on the prices of grains and oilseeds was estimation of the parameters of the modified M.2 model specification of the following form: M.3Δyi,h=α0+α11,hLF(xt)Δxj,t+α12,hLF(xt)Δxj,t−1+α21,hH(1−F(xt))Δxj,t+α22,hH(1−F(xt))Δxj,t−1+ month +εt where: F(xt)=exp(−γ⋆xt)1+exp(−γ⋆xt) , γ > 0 is the so-called “smooth transition function” proposed by Auerbach and Gorodnichenko [2012], for estimating the parameters of the function of the economy’s response to fiscal shocks in various macroeconomic conditions. Function F uses the logistic transformation of the standardized parameter x, which in our application illustrates the conditions on the domestic currency market identified by the level or volatility of the EUR/PLN.⁵ The estimated parameters of the model have the following interpretations:

Sum of parameters α11,hL+α12,hL means a cumulative change in the prices of the analyzed commodity in PLN in response to a change in the price of futures contracts in EUR in the horizon of h months in conditions of a low EUR/PLN or low EUR/PLN volatility;

To assess whether the reaction functions differ significantly depending on the situation on the currency market, the standard Student’s t-test was used to verify the equality of the following linear combination of parameters: α11,hL+α12,hL=α11,hH+α12,hH . The null hypothesis in this test is that the sums of the two parameters do not differ from each other. The sums of both parameters and the significance level of the linear restriction test for each projection horizon are presented in the tables with the estimation results.

The conducted analysis shows as follows (Table 4): (a)

Wheat: The impact of futures changes on the domestic wheat prices varies depending on the level and volatility of EUR/PLN. In the conditions of PLN depreciation, an increase in wheat futures prices results in stronger fluctuations of domestic wheat prices than in the conditions of PLN appreciation. These differences are statistically significant up to the fourth month from the moment of changes in futures prices. The increased volatility of EUR/PLN also contributes to strengthening of the reaction of domestic prices to changes in futures prices, with this effect being statistically significant only in the month in which changes in futures prices occur. These results therefore indicate that in conditions of significant depreciation and (often accompanied by) high volatility of the exchange rate, domestic prices are adjusted to a greater extent to world prices, probably due to the potentially higher costs of not adjusting, than in the case of a strong and stable zloty exchange rate.