Retail Food Price Modelling Project Report

Abstract

This research was undertaken on behalf of Defra under Invitation to Tender Number 24580 to provide further research into modelling UK retail food price inflation. It follows from previously-commissioned work on retail food inflation delivered to Defra in November 2011 comprising the majority of the current research team. In the previous research, the team developed an econometric model to estimate the impact of world commodity prices on UK food inflation while accounting for a range of wider factors that may impact on world and domestic retail food prices and, more specifically, the magnitude of the world-retail price transmission effects. The estimates from the model were then used to forecast the impact of developments on world commodity markets as well as other factors influencing price transmission such as exchange rates and oil prices on UK retail food inflation. The context for the previous research on forecasting retail food inflation in the UK was the exposure to shocks emanating from world markets, in particular following the world price ‘spikes’ of 2007-2008 and 2011. This was a period of considerable price volatility on world markets involving not only agricultural commodities but also oil, both of which were reflected in volatile domestic retail food inflation. The issue of high and volatile food inflation was not confined to the UK, though the experience of food inflation in the UK was more intense than in many other European Union (EU) countries. Since then, however, the challenges facing the UK food system and the exposure of UK consumers to these challenges have changed, particularly following the UK’s decision to exit the European Union. Given the uncertainty that will be involved in the UK’s departure from the EU and the range of alternative trading arrangements that may replace the current trade regime, it is desirable to renew the effort in understanding how the UK food sector (from the farm level through to final consumers) will be affected by these new trade arrangements. To address these issues, the framework initially conceived and delivered under the previous research contract had to be significantly revised. In addition to accommodating more recent data in the econometric model, an innovative feature of the revised food inflation model is the creation of new price indices to reflect specifically what the UK imports and from where. Employing a readily-available world commodity price index, as done in the previous research, is not fit for this specific purpose for two reasons. First, as a portmanteau measure of prices on world commodity markets, it does not reflect the price of what the UK actually imports. Specifically, it ignores that the UK purchases a significant quantity of its food and agricultural products from the EU which are subject to the EU’s common trade policy and hence prices differ from those on world markets. Second, since the UK’s exit from the European Union is likely to involve fundamental changes to international trade measures with potentially both EU and non-EU countries, import prices are likely to change. Hence, it will be necessary to reflect agricultural and food import prices from different sources, including changes to any tariff and non-tariff measures that are applied. Furthermore, changes to the geographical source of UK food imports will also have implications for other data series that impact on domestic food prices, most obviously exchange rates. In particular, it necessitated the creation of an effective exchange rate that reflects the composition of agricultural and food trade that is potentially different from that applicable in the past. These newly-created agricultural-food price and appropriately-weighted effective exchange rate indices form key inputs into the revised framework. There are also a number of other new features to the current research. In particular, since a large part of the UK’s agricultural and food trade relates to processed food products, we create bespoke price indices that reflect the importance of products produced at ii different stages of the food chain. Specifically, we develop models that are based on (a) prices (and effective exchange rates) of agricultural commodities only and (b) a model that includes processed and manufactured food as well as raw agricultural commodity prices and the appropriately weighted exchange rates. While the ‘all-product’ model is our primary focus, the ‘agriculture-only’ model is useful because it facilitates a direct comparison of UK import prices with the prices of agricultural products on international markets. The distinction is important with implications for food price transmission in the UK. In this report, we present the results and insights from this new research. After setting out the context and objectives of the current project, we present an update to the previous food price inflation model (which we label ‘Defra I’) with more recent data to assess its current validity. Following this, we present the derivation of the new price and effective exchange rate indices which form the main new inputs into the revised modelling (labelled ‘Defra II’). It should be noted that, due to the nature of the new data we are working with, and the intention of differentiating between trade with EU and non-EU countries, the specification of the Defra II econometric model differs from the Defra I specification. After an explanation of its structure, estimates of the new model are presented. The results from the econometric model confirm the important role played by domestic factors (including non-agricultural costs (such as labour and energy) and domestic agricultural output prices) as well as international factors (import prices and exchange rates) in determining UK retail food inflation. A major output of the project is the creation of a bespoke Excel-based modelling tool referred to as the ‘Scenario Tool Exeter Food Inflation’ (STEFI). This tool uses the econometric outputs developed in Defra II to deliver estimated effects of alternative post-EU exit trade scenarios on retail food prices. STEFI is user-friendly and has been developed specifically for use within Defra to calculate the dynamic effects of a wide range of scenarios with the option to build-up sequentially the effects in combination with factors such as non-tariff barriers and exchange rates. Finally, in the last two sections of the report, results of two special studies are presented. The first is an analysis of retail food price inflation for different income deciles of the UK population. The second investigates whether the entry of discounters into the UK retail food sector since 2010 may have altered the transmission of agricultural and processed food prices. Our investigation of these issues are tentative in nature and suggest the need for further research to identify the impact of new trade arrangements on different income groups and how the changing structure of the food sector may impact on the transmission of prices through the food supply chain to retail. Taken together, the revised retail food inflation model that is contained in this report represents a substantial development of previous modelling and provides Defra with the flexibility to assess the retail food price effects that may come about as the UK’s trade arrangements with the European Union change and trade arrangements with non-European Union countries develop.