Determinants of food price in Turkey: A Structural VAR approach

Abstract

The rise in food prices in Turkey in the post-Covid 19 period is extremely higher than the increase in the consumer price index and the social dimension of food price. This issue prompted researchers and policy makers to examine these topics. The fact that negativities which have been observed worldwide in an emerging process are making a difference also reinforces the need for more intensive investigation of food prices or policies to be followed. This trend aims to identify the global and macroeconomic factors that cause the rise in food prices in the Turkish economy. According to this purpose, In order to measure the global impact, oil prices, the global fertilizer price index, and the world food price index were included in the study. The M2 exchange rate and money supply variables were used to measure the macroeconomic impact. In this context, monthly data for the period 2003:01–2022:03 were used in the analysis using the SVAR model. Result of the analysis indicated that global and macroeconomic factors had an impact on food prices in the period under consideration. As a result of the variance decomposition, it was observed that the food price after its own collisions was more reactive with the exchange rate, respectively.

Keywords:

• Turkey
• inflation
• food price
• SVAR

1. Introduction

Food inflation is among the most important goals of policy makers regardless of the level of development in countries. This situation can be related to the fact that the concept in question plays a major role in the continuation of human life. The necessity of planning for many different factors such as the sustainability of social life, access to food resources, price, demand, structural problems, and economic balance increases interest in the subject. In addition, when the subject is dealt with in the context of economic literature, it would be appropriate to define inflation first and then food price inflation. In general, inflation can be defined as continuous increases in the general level of prices. On the other hand, food inflation can be expressed as the increase in food items included in the CPI is greater than the increase in the general CPI (Yavuz, 2021, p. 15). Policy makers, who aim to prevent temporary price changes in the general inflation target, implement a sustainable price policy, and take the necessary steps to balance supply and demand, have similar goals in terms of food inflation. It will be important to consider the variations, which are caused by climatic conditions that have an impact on agricultural production due to their characteristic structure of food prices, reduced price elasticity of demand, changes in supply and demand resulting from different influences. It has been established that there is a structure of food prices which can be affected by international and national factors.

When the data of the Food and Agriculture Organization (FAO) is examined, it turns out that the global crisis of 2008, the COVID-19 epidemic and finally the Ukraine-Russia war had an impact on food prices. As an indication of this situation, it can be seen in Figure 1 that the global food price index reached its highest value in 2011 after the global crisis of 2008. After the summit, the index, which had a declining trend in the period up to 2020, entered an upward trend in 2021 with the impact of the pandemic that surrounded the whole world in 2020, and with the Ukraine-Russia war that started in the second. In the month of 2022, the value of the Food Price Index peaked at 159.7 in March. In this context, it can be emphasized once again that the concept in question has a global and complex structure. Below is a graph generated from FAO Food Price Index data.

Figure 1. Food price index, FAO, www.fao.org.

When the causes of fluctuations experienced on a global scale are examined, it is seen that weather change, economic problems and the economic policies implemented in the said period come to the fore. In this context, in addition to the slowdown in the world economy caused by the 2008 crisis, increasing oil prices and the slowdown in agricultural production had a negative impact on food prices on a global scale. In addition to these, countries are searching for new policies in response to crisis as a result of the current lack of conventional policy aimed at solving crises and an adverse impact on supply demand equilibrium is already caused by their pursuit of protectionism. In addition to these negativities experienced on a macro scale, the drought problem that has arisen due to weather change has also brought the global food price to the top (Kutlu, 2021, p. 585). It is seen in Figure 1 that the improvement in climatic conditions and the decrease in economic fluctuations in the following period have reduced the global food price to the level of 2009. However, the onset of the pandemic in 2020 and the implementation of many restrictive measures in the context of combating the pandemic caused the downward trend in food prices to reverse. Among these measures, bans on food exports, which are carried out in order to ensure food safety and meet the need, can be given as an example. Another factor in the upward trend in the global food price index is the Ukraine-Russia war that broke out in February 2022. In this context, it is known that Russia is one of the largest grain exporting countries and the negative consequences of the war could not be resolved through diplomacy in the early stages of the war, which had a negative impact on food prices.

A similar effect was felt on a global scale before, with Russia’s decision to ban grain exports in 2011. Considering the transferred factors, it is possible to collect the factors affecting the food price under several headings. These headings can be considered as global impact, macroeconomic indicators and economic policies, and weather change, respectively. In the global impact title; While there are titles such as global food price, international trade, international food supply chain, war, epidemic, in the title of economy; Population growth, rapidly changing demand structure, monetary policy, balance policies to ensure market balance, foreign trade policy, exchange rate, measures against speculative movements can be examined. The third subheading can be taken into account regarding pollution of the environment, worldwide warming and climate change.

It is thought that these developments on a global scale also have an impact on food prices in Turkey. It can be thought that the crisis that occurred in 2008 and the increase in food prices on a global scale after it caused an increase in food prices in Turkey. In this context, the global crisis, the increasing oil price, the Russian crisis in 2016, the climatic conditions, the slowdown in supply with the pandemic, the sudden increase in demand in the market, the disruptions in the supply chain on a global scale explains the increase in food prices in Turkey (Yavuz, 2021, p. 16). After 2018, the upward trend in the exchange rate and supply shocks increased the input costs and caused an upward pressure on the food price. In the following process, the pandemic’s impact on the whole world and the Ukraine-Russia war that started in 2022 also had an accelerating effect on the upward movement in food prices. As an indicator of this situation, the food price index increased from 550 points in March 2020, when the first fatal case was seen in Turkey, to 1101 points by March 2022. The upward trend in the food price index, which was below 500 points compared to the pre-pandemic period, is shown in Figure 2.

Figure 2. Food price index, www.tcmb.gov.tr.

The determinants of Turkish food prices will be followed through this study on the basis of a statistical approach. The importance of this study is enhanced by the fact that it constitutes a global concept and must be viewed from a broader perspective, with particular relevance for maintaining human life as well as its contribution to setting up more sustainable price policies. On the other hand, when the studies on food inflation and the theoretical infrastructure are examined, it has been concluded that many different variables have an effect on food inflation. In the theoretical framework, it has been seen that the variables in question can be grouped under three headings (Kutlu, 2021). These can be listed as supply-side, demand-side and macroeconomic variables, respectively. In this context, the study focuses on the impact of monetary policy decisions and macroeconomic indicators, which are thought to have an impact on a global scale, in order to measure the macroeconomic impact. As it is known, central banks, which pursue the main objective of price stability, have an important role in ensuring stability in the market balance with the help of the policies they implement. In this context, considering the theoretical framework, it is known that there are many different approaches. In this study, the monetarist approach that establishes a relationship between money supply and inflation within the framework of monetary policy will be included. In the approach pioneered by Milton Friedman, money supply is pointed out as the source of fluctuations in the economy and the resulting inflationary situation. It is also expressed by this view that an irregular money supply expansion will cause instability. From this point of view, it is emphasized that money supply should be changed at certain rates according to economic conditions in order to ensure price stability by establishing a link between money supply and inflation. Based on the theory in question, emphasizing that the most effective policy is monetary policy, the effect of monetary policies implemented by the Central Bank in Turkey on food inflation will be analyzed. In the light of previous studies, econometric analysis will be used by including global-scale or national variables that are thought to have an effect on food price.

1.1. Literature review

This section presents studies on food prices that have an economic and social dimension. The literature review has shown that the variables considered to have an effect on food prices differ between countries or groups of countries chosen, and various results are produced by means of a preferred analysis model as well as historical range. This study necessitates the inclusion of international studies and national studies, respectively. For example, Baek and Koo (2010) examined the factors which influence food prices in the United States by using cointegration analysis as well as confirming that variables such as agricultural commodity price and exchange rates have a major role to play in determining food prices. They have also claimed that in the short and long term, these variables are explained by movements in food prices.

In addition, Lambert and Miljkovic (2010) conducted another study on food price in the case of America. In their study, using time series analysis, they emphasized that the main determinants of food price were agricultural prices and wages rather than energy and consumer incomes. Roache (2010) used the spline-GARCH model in his study to analyze the volatility in food prices in the United States. In his study, which aims to measure low frequency volatility, he stated that there is a positive correlation between different commodities. He also concluded that since the mid-1990s, exchange rate and inflation have an important role in explaining volatility.

Another study examining the factors affecting food price belongs to Rehman and Khan (2015). They used vector error correction and Johansen cointegration tests in their study examining food inflation in Pakistan between 1990 and 2013. As a result of the econometric model, they concluded that indirect taxes and food exports have a statistically positive effect on food inflation. They also found a negative relationship between GDP and food inflation. Another study on food inflation in Pakistan is by Khan and And Gill (2010). They used exchange rate, budget deficit and wheat support price as explanatory variables in their models in which they included different inflation values in the 1970–2007 period. As a result of the analysis, they stated that the wheat support price had no effect on inflation. Another study analyzing high food inflation in Pakistan conducted by Joiya and Shahzad (2013). In their study, they used cointegration and error correction model and included GDP, food exports, food imports and the total amount of credit given to the agricultural sector as explanatory variables. As a result of the analysis, they emphasized that all four variables included in the model had a statistically significant effect on food price in both the long and short term.

A study conducted on food price in Pakistan by Ahsan et al. (2012). They analyzed the determinants of food prices in their studies using the ARDL model in the period 1970–2010. They stated that the most important variable affecting the food price is the money supply. They also emphasized that the long-term effect of subsidies is low and that the rise in global food prices may suppress food prices. Another study examining the effect of macroeconomic factors on food price is Mushtaq et al. (2011). They analyzed the cointegration relationship between wheat price and macroeconomic indicators in the long run. As a result of the analysis, they concluded that the real money supply, openness, and real exchange rate have a statistically significant effect on the wheat price. Moreover, a study that analyzed the determinants of food inflation conducted by Davidson et al. (2012). They included the global food price, exchange rate, production cost, oil price and wage as explanatory variables, which will have a potential impact in their study for England in the 1990–2010 period. In line with the results, they emphasized that the world raw food price and exchange rate are the major variables affecting the food price. For instance, Ferrucci et al. (2010) analyzed the pass-through of commodity price shocks to food prices in the Eurozone. In their study, in which they used the var model, they emphasized that the commodity price had an effect on the food price.

Khan and Gill (2007) used the least squares method in their study examining the effect of money supply on food price and general inflation. Among their results, they stated that food inflation was affected differently by M1, M2, M3 different money supply definitions. They also stated that the money supply affects the overall CPI more than food inflation. Another study stating that money supply has an effect on food inflation is by Khan and Schimmelpfennig (2006). In their study, they emphasized that, in addition to private sector loans, money supply also has an effect on food inflation. Another similar study is by Khan and Qasim (2014). In their study, they reported that the money supply had a statistical effect on the food price.

In a study specific to Turkey conducted by Kutlu (2021) analyzed the determinants of food prices by using the SVAR model. In this study, the period of 2008–2020 was analyzed. The study emphasized that the global food price, industrial production index and food exports did not have an effect on the food price. He stated that only the exchange rate of the variables included in the model has an effect on the food price. Similarly, Çıplak and Yücel (2004) emphasized in their study that the exchange rate has a significant effect on food price. Finally, in another study conducted specifically for Turkey, Başkaya et al. (2008) emphasized that climatic conditions and global food price have an effect on food price.

1.2. Methodology

The SVAR model was developed on the criticism of the standard VAR models. In this context, the main criticism is that the impulse-response functions may not be explanatory because the standard VAR model error terms structure causes the effect of shocks to be ignored in cases where it is not compatible with the literature, and the order of the variables included in the model is important due to the fact that the VAR models work based on the Cholesky decomposition. From this point of view, Sims (1986) and Bernanke (1986) developed the SVAR model, which allows to eliminate the criticisms and to put appropriate constraints on the theory. The SVAR model can be expressed as follows.

(1)

In Equation 1(1) , While yt, lnoil, lnfer, lngfi, lnexc, lnm2, lntrfood represent the vector of endogenous variables, e_(t) represents the 6-dimensional vector of unrelated structural shocks. The reduced VAR model can be expressed as follows.

(2)

In Equation 2(2) , ε_t represents the vector of reduced error terms, . Structural shocks are determined using the zero constraint in matrix B. The restrictions can be expressed as follows.

2.

2.1. Data set and model

In this part of the study, an econometric model will be established for the determinants of food inflation in Turkey. In this context, monthly data for the period of January 2003 - March 2022 were used for the indicators of food inflation in the Turkish economy. As the dependent variable, the food and non-alcoholic beverages index, which is the sub-index of the CPI index, was preferred. Nominal exchange rate, M2 money supply, global food price index, global fertilizer price index and oil price index were chosen as independent variables. The data of the variables were obtained from the IMF and TCMB databases. After the variables were adjusted for seasonality with the help of the Census X12 test, the analysis was continued by taking their natural logarithms. The abbreviations of the variables used in the study and the sources from which they were obtained are given in Table 1.

Table 1. Variables and sources

Variables	Source
LNTRFOOD (food and non-alcoholic beverages in Turkey)	CRBT
LNGFI (global food price index)	IMF
LNOİL (global oil price index)	IMF
LNFER (global fertilizer index)	IMF
LNEXC (nominal Exchange rate in Turkey)	IMF
LNM2 (M2 money supply in Turkey)	IMF

When the matrix in the above equation is examined, it expresses the simultaneous response of the first variable in the first row of the matrix to the shocks of other variables, while the first column shows the effect of the shock in the first variable on other variables. From this point of view, it is assumed that the LNOIL variable only responds simultaneously to its own shocks, in line with the literature. In other words, LNOIL variable is not affected by shocks in other variables.

On the other hand, it is assumed that rising oil prices will also have an impact on fertiliser and global food prices taking into account energy which is a major input for production. Therefore, the effect of the LNOIL shock on the LNFER and LNGFI variables has been released. Furthermore, it has been considered that oil prices also have a influence on the exchange rate and money supply given their effect on several economic indicators as well as Central Bank monetary policy aimed at ensuring price stability. Consequently, the effects on exchange rates and money supply due to LNOIL shocks are also exposed in accordance with literature.

Finally, since all the variables included in the model are thought to affect the LNTRFOOD variable, which represents the food and nonalcoholic beverages index in Turkey, it is assumed that it will react simultaneously to the matrix created.

2.2. Analysis results

In time series models, firstly, the analysis is started with unit root test in order to prevent spurious relationship or misleading estimations. In this context, the existence of unit root in the variables included in the model was analyzed with Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) unit root tests. When the results of both tests were examined, it was concluded that the variables included in the model were not stationary at their levels. In the test results in the second stage, which was made by taking the differences of the variables, it was seen that the variables rejected the main hypothesis of “the series contains unit root”.

After it was seen that the variables included in the model became stationary by taking the first difference, information criteria were used to determine the appropriate lag for the VAR model. According to LR, FPE, AIC and HQ information criteria, the appropriate lag length was determined as 2. In the established model, the results of impulse-response (with Cholesky ordering) and variance decomposition analyzes were used to interpret the relationship between the variables. The response of the LNTRFOOD dependent variable to the shocks occurring in the variables added to the model as an independent variable is interpreted in the action-response graphs. In variance decomposition, it is analyzed to reveal how much of the change in one of the variables is due to other variables.

The first graph of Chart 1 shows the reaction of the food price index in Turkey to the one standard deviation shock in the oil price. It is observed that food prices reacted positively to oil shocks with an increasing trend, especially after the second period. It can be stated that this reaction, which has an effect on food prices. Baek and Koo (2010), Nazlioglu and Soytas (2011) and Sujıthan et al. (2014) are similar to their studies. The second graph shows the reaction of food prices to the shock in global fertilizer prices. It is seen that the food price responded to the shock that occurred in the global fertilizer price index (LNFER) from the first period. In other words, it has been concluded that food prices are affected by fertilizer prices. The emergence of this situation can be attributed to the fact that phosphate exporting countries impose export restrictions and the increase in natural gas prices has a significant upward effect on fertilizer prices (World Bank, 2022, 9). The third chart takes the reaction of the LNTRFOOD variable to shock in the Global Food Prices GFI index and reproduces it. In accordance with the results, it was found that LNTRFOOD reacted in a positive way to LNGFI shock during the 1st period and remained so for the following two periods. This result is similar to the studies of Davidson et al. (2012) and Ferrucci (2010), which emphasize the impact of global commodity prices on food prices. In the fourth chart showing the reaction of the food price index against the shock in the exchange rate, it is seen that the food price reacted positively. This response is similar to Akanni (2020), Gilbert (2010), Kutlu (2021), who emphasized that the exchange rate has a significant effect on food price. The response to the shock in the money variable M2, which is included in the model to examine whether the increases in the food price index are due to demand, is given in the fifth chart. It is seen that he gave a statistically significant response to the shock in question. From this point of view, it can be stated that the increase in money supply has a positive effect on the food price index. This result is consistent with the studies of Bordo (1980) and Chambers (1984). The last chart shows the reaction of the food index to its own shocks. The effect of the variable’s reaction to the shock that occurs in its own delay decreases in the following periods.

Chart 1. Impulse response graph.

In order to find out how much change in one of the variables has been caused by another, variance decomposition analysis results are also present as a result of switching impulse response graphics from Structural VAR model. Table 3 shows the results of the variance decomposition analysis of the food price index. According to the results in the table, it is seen that the biggest source of the change in the food price index in the first period is its own shocks. In the first period, it was observed that 84.3% of the change was caused by itself and 13.6% by the LNEXC variable. The effect of other variables in the first period was relatively less compared to these two variables. In the following periods, it is seen that the effect of the variables on the food price index changes. From this point of view, the impact of the food price index’s own shocks gradually diminishes in the following period, as expected. It is observed that the effect of the exchange rate variable increases in the following period. It was shown that the LNEXC and LNTRFOOD variables had a significant influence on the change of the dependent variable at levels of 60.7% and 20.1% during the 10th period. Besides those results, the LNEXC variable is a major factor influencing changes in food price index after its own shocks. On the other hand, the effects of LNFER, LNGFI, LNOIL and LNM2 variables increased in the subsequent periods compared to the first period, among the results obtained (see Table 2).

Table 2. Unit root test results

	I(0)		I(1)
	Intercept	Int and Trend	Intercept	Int and Trend
ADF Unit Root Test
LNOIL	−2,4750	−2,5651	−11,1323*	−11,1035*
LNFER	−0,9755	−1,1319	−9,5254*	−9,4952*
LNGFI	−1,3836	−1,8977	−10,6080*	−10,5823*
LNEXC	3,3407	−0,7559	−12,3653*	−13,0330*
LNM2	−1,2568	−2,2127	−15,2169*	−15,2339*
LNTRFOOD	3,5324	4,7214	−10,8928*	−11,5155*
PP Unit Root Test
LNOIL	−2,2792	−2,3345	−10,8116*	−10,7852*
LNFER	−1,6453	−1,7172	−8,5253*	−8,4944*
LNGFI	−1,3929	−1,8899	−10,5875*	−10,5617*
LNEXC	3,2363	−0,6117	−12,3537*	−12,8878*
LNM2	−1,1933	−2,2512	−15,2992*	−15,3075*
LNTRFOOD	3,8112	4,3025	−11,0520*	−11,5435*

Notes: The values in the table are t statistical values. In the PP unit root test, Bartlett Kernel and Bandwidth option Newey-West were used as the spectral estimation method. In the ADF unit root test, the lag length criterion and the Schwarz information criterion were chosen. * indicates the 1% level. Authors’ own estimations.

As a result of the findings, it has been reached that the variable that has a statistically significant effect on the food price index is the exchange rate. This result once again proves the importance of the exchange rate on macroeconomic indicators. It also shows that different sectors should be taken into account in preferred monetary policies in order to create foreign exchange savings or foreign trade balance. Considering the characteristic structure of the Turkish economy, the fact that some of the food production is dependent on imports highlights the pass-through from exchange rates to food prices. As an indicator of this situation, while the value of the fertilizer and soil improvers index, which is included in the agricultural input price index in Turkey, was 252 in July 2021, it was 844 in July 2022 (TURKSTAT, 2022). In addition, the amount of fertilizer imports realized was approximately 1.8 billion dollars (TURKSTAT, 2022). On the other hand, it is thought that the cost increases that occur with the increasing exchange rate will cause deterioration in the supply-demand balance. As an indication of this situation, it can be said that rising costs will lead to a reduction in supply, which will lead to a vicious circle in the future and unexpected increases in food prices, which will lead to a deterioration in the supply demand balance.

It can be said that the said exchange rate increases have an impact on demand as well as on supply. Increases in the prices of food products, which have an important place in the survival of people, may have a negative impact on food safety and access of low-income groups. In this context, since the exchange rate is important for both parties, it becomes inevitable for the central banks, whose main objective is to ensure price stability, to intervene in order for this situation to occur. At this point, it is thought that it would be useful to specify a subject. These issues, respectively, may have a negative impact on the market balance, such as supply chain disruptions and the desire to make excessive profits. On the other hand, as can be seen in Table 3, the effect of exchange rate on food prices starts to show its effect especially after three periods. In this situation, it is important for policy makers to use the necessary tools, starting with a strict verbal intervention in the first 3 months instead of a wait and see strategy.

Table 3. Variance decomposition

Period	S.E.	LNOIL	LNFER	LNGFI	LNEXC	LNM2	LNTRFOOD
1	0,085148	0,027541	0,529262	0,45982	13,67866	0,91872	84,386
2	0,136273	0,013288	0,476376	1,311559	25,16272	0,613853	72,42221
3	0,173497	0,081295	0,760004	1,588471	35,12653	0,96497	61,47873
4	0,20084	0,364623	1,440059	1,701248	43,03175	1,47234	51,98998
5	0,221413	0,873928	2,436485	1,828411	48,91235	1,964002	43,98483
6	0,237365	1,540378	3,599005	2,029979	53,16071	2,3715	37,29843
7	0,25011	2,277941	4,801216	2,318205	56,18668	2,679491	31,73647
8	0,260568	3,017754	5,965244	2,686715	58,31668	2,89389	27,11972
9	0,26934	3,716007	7,051716	3,12361	59,79201	3,027674	23,28898
10	0,27683	4,349754	8,044832	3,616792	60,78782	3,09522	20,10558

Source: Authors’ own estimations.

3. Conclusion

In this study, we aimed to determine the global and macro-economic variables that are thought to have an effect on the food price in the Turkish sample. As a result of the literature review, oil price, global fertilizer price, global food price variables were included in the model in order to measure global effects, while exchange rate and M2 money supply variables were included as macroeconomic indicators. In the analysis using the SVAR model, the effect levels of the mentioned variables were included with the help of impulse-response graphs and variance decomposition table. As a result of the analysis, it was concluded that the variables thought to have an effect on the food price had an effect on the food price. From this point of view, it has been concluded that the fluctuations in food prices are affected by both the global developments and the changes in macroeconomic indicators. Given the information in the variance decomposition chart, it is shown that food prices have the largest effects on exchange rates as well as global fertilizer price except for their own shocks. In this context, it must be stressed that the exchange rate has been a major determinant of food prices during the period under consideration. It is also among the results obtained that the food price reacts to the oil price, the global food price and the money supply, respectively. As a result of the findings, it is thought that the exchange rate and global effects have a determining effect on the food price in the period discussed. At this point, it is important to ensure exchange rate stability in the policies to be implemented in order to ensure price stability. On the other hand, it can be stated that the weak reaction of the food price to the money supply should be considered in terms of permanent price stability. To put it another way, this unsatisfactory response shows that the impact of expansionary policies to be implemented for economic growth acceleration and elimination of the devastating effects of inflation on food prices should not be overlooked.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

İ̇brahim Orkun Oral

İbrahim Orkun Oral, Department of Finance and Banking, Faculty of Applied Sciences, Bilecik Sheikh Edebali University Faculty of Applied Sciences, Istanbul, Turkey, Yeşilkent street 11300 Bozüyük Bilecik, E-mail: [email protected] TEL-FAX : [phone] Avni Çakıcı, Department of Food Engineering, Faculty of Engineering, İstanbul Aydın University, Istanbul, Turkey, E-mail: HYPERLINK “mailto:[email protected]” [email protected].

Avni Çakıcı

Fatih Yıldız, Department of Food Engineering, Fatih Yildiz, Food Engineering and Biotechnology, Middle East Technical University, Ankara, Turkey, E-mail: HYPERLINK “mailto:[email protected]” [email protected].

Fatih Yıldız

Mohannad Alayoubi, Department of Food Engineering, Faculty of Engineering, İstanbul Aydın University, Istanbul, Turkey, E-mail: [email protected].