Can urban coffee consumption help predict US inflation?

• Author: Idris A. Adediran,Raymond Swaray,Afees A. Salisu
• Date: 01 November 2019
• DOI: http://doi.org/10.1002/for.2589
• Published date: 01 November 2019

RESEARCH ARTICLE

Can urban coffee consumption help predict US inflation?

Afees A. Salisu

1,2

| Raymond Swaray

3

| Idris A. Adediran

4,5

1

Department for Management of Science

and Technology Development, Ton Duc

Thang University, Ho Chi Minh City,

Vietnam

2

Faculty of Business Administration, Ton

Duc Thang University, Ho Chi Minh City,

Vietnam

3

Economics Subject Group, University of

Hull Business School, Hull, UK

4

Centre for Econometric and Allied

Research, University of Ibadan, Ibadan,

Nigeria

5

Department of Economics, Obafemi

Awolowo University, Ife, Nigeria

Correspondence

Raymond Swaray, Economics Subject

Group, University of Hull Business

School, Cottingham Road, Hull, UK

Email: r.swaray@hull.ac.uk

Abstract

Motivated by the importance of coffee to Americans and the significance of the

coffee subsector to the US economy, we pursue three notable innovations.

First, we augment the traditional Phillips curve model with the coffee price

as a predictor, and show that the resulting model outperforms the traditional

variant in both in‐sample and out‐of‐sample predictability of US inflation. Sec-

ond, we demonstrate the need to account for the inherent statistical features of

predictors such as persistence, endogeneity, and conditional heteroskedasticity

effects when dealing with US inflation. Consequently, we offer robust illustra-

tions to show that the choice of estimator matters for improved US inflation

forecasts. Third, the proposed augmented Phillips curve also outperforms time

series models such as autoregressive integrated moving average and the frac-

tionally integrated version for both in‐sample and out‐of‐sample forecasts.

Our results show that augmenting the traditional Phillips curve with the urban

coffee price will produce better forecast results for US inflation only when the

statistical effects are captured in the estimation process. Our results are robust

to alternative measures of inflation, different data frequencies, higher order

moments, multiple data samples and multiple forecast horizons.

KEYWORDS

coffee price, forecastevaluation, inflation forecasts, Phillips curve, USA

1|INTRODUCTION

Over the past decades, the link between prices of com-

modities (fuel and nonfuel) and macroeconomic/

financial indicators such as economic activity, inflation,

interest rates, exchange rate, stock price, and money has

been widely established in the literature (see, e.g., Baillie,

1989; Blomberg & Harris, 1995; Boughton & Branson,

1988; Boughton, Branson, & Muttardy, 1989; Browne &

Cronin, 2010; Ciner, 2011; Cody & Mills, 1991; Frankel,

2013; Furlong & Ingenito, 1996; Garner, 1989; Marquis

& Cunningham, 1990; Pecchenino, 1992; Sephton, 1991;

Webb, 1988). Equally, a number of studies have singled

out the role of commodity prices as an indicator of con-

sumer inflation (see Bloch, Dockery, & Sapsford, 2004;

Chen, Turnovsky, & Zivot, 2014; Davidson, Halunga,

Lloyd, McCorriston, & Morgan, 2016; Ferrucci, Jiménez‐

Rodríguez, & Onorante, 2010; Gelos & Ustyugova, 2016;

Gilbert, 1990; Gómez, González, & Melo, 2012; Kagraoka,

2015; Kyrtsou & Labys, 2006; Richards & Pofahl, 2009).

These studies are largely motivated by the implications

of commodity prices for the effectiveness of policy deci-

sions (especially monetary policy). Since commodity

prices precede inflation, information about their varia-

tions can be exploited by monetary authorities in policy

decisions. This conclusion is particularly valid for the

USA, where the consumption of commodities is ranked

the largest after China, and therefore inflation dynamics

are more likely to be closely related to movements in

commodity prices.

Studies on the commodity price–US inflation nexus

tend either to use crude oil (a fuel commodity) or gold

Received: 23 April 2018 Revised: 12 January 2019 Accepted: 8 March 2019

DOI: 10.1002/for.2589

Journal of Forecasting. 2019; :649–668. © 2019 John Wiley & Sons, Ltd.wileyonlinelibrary.com/journal/for

649

38

(a metal commodity which has historical links to the US

dollar and other major currencies), or a broad index of

commodity prices (see Edelstein & Kilian, 2009; Van

Hoang, Lahiani, & Heller, 2016; Hooker, 2002; Kilian,

2008; Mahdavi & Zhou, 1997; Valcarcel & Wohar, 2013;

Salisu & Isah, 2018). So far, no study has examined the role

of coffee alone in predicting US inflation. We argue that

there are several motivations for this special consideration

of coffee consumption in predicting inflation in the USA.

First, statistics from the US Department of Agriculture

show the USA as the largest single country consumer of

coffee, importing about 2.5 million 60 kg bags of coffee

beans in December 2017 (US Department of Agriculture,

2017).

1

Unlike other commodities such as crude oil, of

which the USA is a major producer, or gold, coffee is exclu-

sively an import to the USA. Coffee as a beverage has

gained increasing popularity among younger adults in

the USA and Europe, and coffee drinking has been linked

to beneficial effects on health and longevity in newspaper

reports.

2

A report by the Financial Times (2018) attributes

the recent growth in US coffee consumption to “café cul-

ture among millennial”coffee drinkers.

Second, at the micro level, coffee is the most widely

consumed beverage by households in the USA. The Min-

istry of Agriculture and Agri‐Food Canada (2013)

highlighted coffee as a daily staple beverage in the USA

that is “consumed across all incomes, ages, genders and

states”and considered as an “instant energy boost for

those on‐the‐go”(p. 2). A study by Mintel (2012) shows

that 76% of US consumers bought coffee produce to use

at home. This reality has implications for inflation

through the exchange rate pass‐through. Thus an unan-

ticipated change in the coffee price could put pressure

on consumer price inflation, given the high consumer

demand nature of the US economy.

Third, US corporations dominate coffee retailing and

marketing around the world. Five out of the top ten spe-

cialist coffee house chains in the world are US‐registered

firms. In particular, Starbucks Corporation (a Seattle‐

based firm) alone has sales revenue which was more than

treble the revenue of the remaining nine combined in

2015 (Euromonitor International, 2016).

3

It has never

been easier to access a cup of coffee at home, in offices,

or in hotel rooms with new, improved coffee makers

and coffee pod machines.

4

A recent report by

Euromonitor International estimates total global spend-

ing on coffee at $180–200 billion per year, and US firms

are at the forefront of consolidation and business model

innovation in the worldwide coffee market. The global

coffee industry has benefited from remarkable innova-

tions in product, process, and business models over the

past decade, which mirror wider innovations in the

USA and the world economy. In addition, the recent

acquisition of UK‐based Costa Coffee (the second largest

coffee chain in the world) by Coca‐Cola and an alliance

established between Starbucks and Nestlé further exem-

plifies US firms' global dominance of the “new age of cof-

fee everywhere”(Euromonitor International, 2018, p. 1).

In this new era of coffee, it is an “American firm”that

has taken “café culture”, which is essentially “Italian,”

to global consumers (Economist, 2016).

Fourth, coffee is a commodity which has a “single final

use,”namely as a “beverage”—a unique characteristic

which distinguishes it from other tradeable commodities.

In addition, coffee “is not combined with any other input

to produce the final good”, though extracting the soluble

component is amenable to a range of manufacturing

methods (Maizels, Bacon, & Mavrotas, 1997, p. 143). Pro-

duction of coffee beans is subject to exogenous supply‐

side shocks, and once produced the beans are not storable

over prolonged periods without deterioration in quality or

at high storage costs, which contributed to a breakdown

in International Commodity Agreements (Gilbert, 1987,

1996; Swaray, 2007). Consequently, we posit that the cof-

fee price (all‐urban CPI for coffee) in the USA is likely to

hold predictive power for inflation in the USA, which

calls for a rigorous empirical investigation.

Generally, the theoretical propositions for considering

commodity prices in this light are situated around the

theory of overshooting commodity prices, found in the

works of Frankel (1986). The commodity markets reflect

all available supply and demand shocks in the economy

and these shocks (including shocks to aggregate demand,

wealth, aggregate investment, technological innovation,

harvest failures and labor shortages, among others) give

room for unanticipated inflation arising from the market

(see Malliaris, 2006). Consequently, in the event of such

unanticipated inflation, when commodity prices rise,

wholesalers and retailers of products respond swiftly by

passing along a certain proportion of the price increase

to consumers (see Richards, Allender, & Hamilton,

1

Note that the European Union as a trading block of 28 countries is the

largest importer of coffee, importing on average 3.84 million 60 kg bags

per month in 2017.

2

See the following articles in the New York Times:“More consensus on

coffee's effect on health than you might think,”May 11, 2015; “Coffee

drinkers may live longer,”July 2, 2018; “Coffee may tame the redness

of rosacea,”October 17, 2018.

3

The top ten global coffee house chains in 2015 were Starbucks (USA),

Costa Coffee (UK), McCafe (USA), Doutor Coffee Shop (Japan), Coffee

Bean & Tea Leaf (USA), Caffe Nero (UK), Tully's (USA), Ediya Espresso

(South Korea), Caribou Coffee (USA), Gloria Jean's Coffees (Australia).

4

Gassmann, Frankenberger, and Csik (2014) questioned whether any-

one would imagine ten years ago that coffee drinkers would buy

Nespresso coffee capsules for €80 per kg.

SALISU ET AL.

650