Household food demand in Turkey: A two-step demand system approach Abdulbaki Bilgic a , Steven T. Yen b,⇑ a Department of Agricultural Economics, College of Agriculture, Atatürk University, Erzurum 25240, Turkey b Department of Agricultural and Resource Economics, The University of Tennessee, Knoxville, TN 37996-4518, USA article info Article history: Received 3 October 2012 Received in revised form 23 August 2013 Accepted 3 September 2013 Keywords: Bootstrap Censoring Food demand LAIDS Turkey abstract Demands for sixteen food products are investigated, using data from the Turkish Household Expenditure Survey. The linear approximate almost ideal demand system (LAIDS) is estimated with Shonkwiler and Yen’s two-step procedure. All own-price elasticities are negative and expenditure elasticities positive. Bread, other cereals, bovine, mutton, giblets, and cheese have high expenditure elasticities. Mutton, bovine, and several other protein-rich products are price elastic. Results suggest a mix of gross substi- tutes and complements, while net substitution is the dominant pattern. Demographic characteristics also play important roles in shaping food demand. The elasticity estimates can inform policy deliberations. Ó 2013 Published by Elsevier Ltd. Introduction Household expenditures in Turkey have experienced significant structural changes in the last few decades. Statistics released by the Turkish Statistical Institute (TUIK) suggest consumption of sta- ple food groups that previously was stable has been on a steady decline for decades (TUIK, 2011). As Turkey has experienced an impressive economic growth since 2002 with a 6.8% average an- nual rate of per-capita GDP growth (IMF, 2009), food expenditure including non-alcoholic beverages as a percentage of total house- hold spending continued to shrink from 26.7% in 2002 to 21.7% in 2011. Percentage of expenditure on transportation has soared, while modest increases are seen in restaurant/hotels and enter- tainment/culture and modest decrease in health, over the last dec- ade (TUIK, 2011). Despite these changes, the proportion of food expenditure remains nearly twice as large as that of the European Union (EU) countries (12.70%) (TUIK, 2011)(Table 1). Between the periods of 1990–1992 and 2006–2008 in Turkey, daily dietary energy supplies (DES) declined slightly from 3590 kcal to 3500 kcal and per capita protein intake from 104.9 g to 99 g, while daily per capita animal protein and fat intakes became an indispensable part of Turkish diet, increasing from 25.7 g to 26.2 g and 89.4 g to 103.1 g, respectively (FAO, 2011). Average daily per capita DES intake fell even further to a hearty 3482 kcal and the average percentage of protein intake was 11% in 2011, while 12% of children suffered from malnourishment and stunts in their growth (FAO, 2011). Annual per capita red meat consumption in Turkey fell from 8.19 kg in 2007, to 6.78 kg in 2008 and 5.73 kg in 2009 (TUIK, 2010). This decline was due mostly to the fact that price of meat more than doubled in the last three years and continued to rise despite government claims of no shortages. On a per capita basis, consumption of red meat in Turkey is about one-fifth of the EU average, while consumption of mutton (sheep) meat exceeds that of the EU level (Bilgic and Yen, 2013). On the other hand, consump- tion of poultry meat increased rapidly, by 171%, from 3.8 kg in 1990 to 14.1 kg per capita in 2005, with an average growth rate of 17.5% per year, exceeding that of red meat (Yalcin, 2006). In spite of the significant increase due mostly to increasing red meat prices, consumption of poultry has fallen behind the goals of the 27 EU member countries (EU27), averaging 15.8 kg per person. Also, annual per capita consumption of dairy products amounts to 121–125 kg which is very low compared to that in EU countries such as Finland (183.9 kg), Sweden (145.5 kg), and Ireland (129.8 kg) in 2006 (International Dairy Federation, 2007). Overall, meat, fish, fruit, and vegetable consumption was estimated to soar from 21.3, 7.0, 105.2 and 226.3 kg per head in 2004 to 24.6, 7.4, 114.0, and 233.2 kg per head in 2013, respectively, whilst milk con- sumption was estimated to increase from 126.6 l in 2004 to 140.8 l per head in 2013 (Deloitte, 2010). The main reasons for increasing consumption in foodstuffs are rising income and changing consumption patterns that have made the Turkish consumers increasingly demanding, driven by the vast array of product choices offered by mass grocery retail outlets throughout the 0306-9192/$ - see front matter Ó 2013 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.foodpol.2013.09.004 ⇑ Corresponding author. Tel.: +1 865 974 7231; fax: +1 865 974 7484. E-mail address: syen@utk.edu (S.T. Yen). URL: http://web.utk.edu/~syen/ (S.T. Yen). Food Policy 43 (2013) 267–277 Contents lists available at ScienceDirect Food Policy journal homepage: www.elsevier.com/locate/foodpol