Modelling Australia’s exports of non-commodity goods and services
Michael Kouparitsas, Linden Luo and Jazmine Smith[1]
Treasury Working Paper[2]
201701
Date created: February 2017
Date modified: February 2017
© Commonwealth of Australia 2017
ISBN 978-1-925504-32-3
This publication is available for your use under a Creative Commons BY Attribution 3. 0 Australia licence, with the exception of the Commonwealth Coat of Arms, the Treasury logo, photographs, images, signatures and where otherwise stated. The full licence terms are available from http://creativecommons. org/licenses/by/3. 0/au/legalcode.
Use of Treasury material under a Creative Commons BY Attribution 3. 0 Australia licence requires you to attribute the work (but not in any way that suggests that the Treasury endorses you or your use of the work).
Treasury material used 'as supplied'
Provided you have not modified or transformed Treasury material in any way including, for example, by changing the Treasury text; calculating percentage changes; graphing or charting data; or deriving new statistics from published Treasury statistics — then Treasury prefers the following attribution:
Source: The Australian Governmentthe Treasury
Derivative material
If you have modified or transformed Treasury material, or derived new material from those of the Treasury in any way, then Treasury prefers the following attribution:
Based on The Australian Government the Treasury data
Use of the Coat of Arms
The terms under which the Coat of Arms can be used are set out on the It’s an Honour website (see www. itsanhonour. gov. au)
Other Uses
Inquiries regarding this licence and any other use of this document are welcome at:
Manager
Communications
The Treasury
Langton Crescent Parkes ACT 2600
Email: medialiaison@treasury. gov. au
Modelling Australia’s exports of non-commodity goods and services
Michael Kouparitsas, Linden Luo and Jazmine Smith
201701
February 2017
Abstract
This paper models both the supply and demand of Australian non-commodity exports. We derive longrun export demand relationships from first principles. On the demand side, the paper finds a relatively low substitution elasticity between Australian exports and a broad basket of foreign produced goods and services. So, for instance, if Australian export prices increase, overseas buyers are less likely to respond by purchasing the same goods and services from foreign competitors, and are instead more likely to respond by reducing their demand for the product – whether Australian or foreign-made. In other words, income effects trump substitution effects. This result is consistent with other Australian studies. On the supply side, our modelling assumes that Australian manufacturing and services exporters are price setters – an assumption consistent with existing literature. This means that if global input costs increase, Australian exporters are able to pass some of that increase onto their customers. Based on this assumption, our modelling suggests that labour costs are a larger contributor to Australia’s non commodity export prices than imported intermediate inputs costs.
JEL Classification Numbers: C22, C53, F17
Keywords: Trade policy; optimal tariff; price elasticity
Michael Kouparitsas, Linden Luo and Jazmine Smith
Macroeconomic Modelling and Policy Division
Macroeconomic Group
The Treasury
Langton Crescent
Parkes ACT 2600
1. Introduction
International trade plays a critical role in fostering improvements in economic welfare, especially in small open economies. It allows economies to grow faster and achieve higher material living standards than they would in the absence of such trade. Theoretical and quantitative trade policy literature has shown that the gains from trade are determined by the price elasticities of export demand and supply. For example, Goldstein and Khan (1985) argue that welfare enhancing policy change in an open economy relies on robust estimates of the price elasticities of export demand and supply. Yet despite its importance, there is little empirical research devoted to estimating export demand and supply elasticities at either the country or global level. This paper adds to this literature by modelling Australia’s exports of non-commodity goods (for example, manufactured goods) and services (for example, education, travel and business services).
This paper is not comparable with all empirical export demand studies. These studies fall into either country specific studies (see, for example: Jilek, Johnson and Taplin, 1993; Senhadji and Montenegro, 1999[3]; Dvornak, Kohler and Menzies, 2005; and Norman, 2006) or global studies (see, for example: Broda and Weinstein, 2008; and Broda, Lamão and Weinstein,2008). Country-specific studies typically model aggregate time-series data, while global studies focus on the welfare effects of increasing product variety by modelling highly disaggregated panel data (that is, data with both a cross-sectional and time dimension). This paper focuses on the former. As such, its results are not directly comparable to modelling that uses more disaggregated panel data.
For completeness we revisit the theoretical export demand framework. Our framework is based on the reasoning that Australia’s exports are the imports of our trade partners. Conditional import demand relationships are derived from representative household/firm-level utilitymaximisation/costminimisation problems for each Australian trading partner. These countryspecific import demand relationships are then summed to form an aggregate export demand equation. Following the broader international trade literature (see, for example, Dixit and Stiglitz,1977), our approach assumes households/firms have both a taste for variety and rival sources of supply. Consistent with other country-specific studies, this long-run theoretical demand framework is augmented by cyclical explanatory variables to form empirical error correction models (that is, shortrun demand models) which are estimated using standard econometric methods.
Over the past 30 years, the imports of Australia and its trading partners have grown at a much faster rate than their respective gross domestic product (GDP). This is generally characterised as rising import penetration. Recent studies that have explored Australian non-commodity export demand (Dvornak et al, 2005; and Norman, 2006) modelled this feature of the data by allowing the foreign income elasticity to be greater than one. But this approach encounters some problems. It is inconsistent with the underlying theoretical framework which implies an income elasticity of one. It is also problematic if the estimated export demand relationship is embedded in a broader macroeconomic model because it is incompatible with balanced growth. In light of this, we deviate from other Australian studies by adhering to the theoretical model by imposing a foreign income elasticity of one and allowing rising foreign import penetration to be captured via a deterministic time trend.
On the demand side, our key finding is that there is relatively low substitutability between noncommodity goods and services that are exported by Australia and a broad basket of foreign produced goods. The main implication of this is that when there is a change in the relative prices of these exports, income effects will dominate substitution effects. So for example, other things being equal, a rise in export prices will reduce purchasing power among the buyers of those exports, leading to a fall in demand for both Australian and competing exports.
The theoretical underpinning of the supply of non-commodity exports is motivated by previous Australian research undertaken by Swift (1998). It suggests Australian exporters have some ability to set prices on the world market. Swift’s (1998) results are consistent with subsequent research by Dvornak et al. (2005) which found that Australia’s manufacturing and services exporters are price setters. Supply of non-commodity exports is modelled via a log-linear price mark-up equation, where the price of the export is a mark-up over nominal unit labour costs (labour cost per unit of output) and intermediate goods costs per unit of output. The final function form is similar to that used by deBrouwer and Ericsson (1998) in modelling aggregate inflation. Again, the long-run theoretical framework is augmented by cyclical explanatory variables to form empirical error correction models (that is, shortrun supply models) which are estimated using standard econometric techniques.
The remainder of this paper is organised as follows: Section 2 derives the theoretical long-run export demand and supply relationships; Section 3 describes the data used in estimating export demand and supply; Section 4 provides details of the econometric method and reports parameter estimates; and Section 5 summarises the analysis and outlines plans for future work.
2. Theory
Long-run demand for exports
For completeness consider a representative consumer living in foreign country i who solves a nested utility maximising problem. At the top of the nest is the choice between different types of consumption goods. Assuming the consumer’s preferences are captured by Dixit and Stiglitz (1977) preferences (that is, a constant elasticity of substitution utility function), the problem can be summarised by the following:
1)
subject to:
2)
where q is the number of varieties, u is the elasticity of substitution between varieties, pi k is a weighting parameter, cit is the ideal/aggregate consumption level, pcit is the ideal/aggregate consumption price, and pcitk the price of consumption good k.
The Lagrangian for the consumer’s maximisation problem is:
3)
where lit is the multiplier associated with aggregate consumption.
The first order conditions for maximisation are:
4)
5)
Combining (4) and (5) implies:
6)
which in turn implies:
7)
Taking logarithms and rearranging implies the following demand relationship for good k:
8)
The aggregate price index can be derived using (5) and (7). Specifically, (5) implies:
9)
while equation (7) implies:
10)
Substituting (9) into (10) implies the aggregate consumption price index is an aggregation of individual goods prices:
11)
Once the consumer has made their decision at the top level, they move to the next level, which involves the choice between domestically produced and imported varieties of the goods subject to the value of aggregate expenditure decided at the previous level. Assuming that Dixit-Stiglitz preferences also apply at this level, the consumer has the following nested maximisation problem:
12)
subject to:
13)
where sk is the elasticity of substitution between domestic and imported varieties of k, q k is a weighting parameter which implies home bias, ditk is the volume of domestically produced good k, pditk is the price of domestically produced k, mitk is the volume of imported good k and pmitk is the price of imported k.
This implies the following demand relationships for domestically produced and imported k:
14)
and price index for good k:
15)
Finally, consumers in country i must choose between imported varieties of good k sourced from different trading partners. Assuming Dixit-Stiglitz preferences also apply at this level yields the following nested maximisation problem:
16)
subject to:
17)
where n is the number of varieties, hk is the elasticity of substitution between imported varieties of k, w kij is a weighting parameter, mijtk is the volume of good k imported from country j and pmijtk is the price of good k imported from country j.
The first order conditions for this maximisation problem imply the following individual import demand equations for country i with respect to imports sourced from country j:
18)
Recognising that one country’s imports are another country’s exports and assuming first stage passthrough[4] of prices implies:
19)
where xkij is the export of good k from j to i, pxjk is the price of country j exports and eij is the countryicountry j exchange rate in terms of country j currency.
This implies the following demand equation for country i with respect to country j exports:
20)
Furthermore, aggregate export demand for country j can be approximated as the share weighted sum of the individual export equations:
21)
22)
where
23)
where skijt is the time t export share for country j. Recognising ln(1+e) is approximately equal to e for small e, implies the following approximate aggregate export demand equation for country j’s good k:
24)
where sijtk = xijtk/xjtk
This equation shows that aggregate demand is driven by a trade-weighted foreign demand index, a tradeweighted exchange rate and a trade-weighted foreign import price index for k.
Assuming the nested decisions have the same elasticity of substitution (s=h) and substituting (15) into (25), the export demand for country j’s good k can be written in terms of aggregate foreign consumption volumes and prices:
25)
This expression can be simplified further by defining trade-weighted foreign consumption, nominal exchange rate and foreign expenditure prices respectively:
26)
and substituting these expressions into (25), which implies:
27)
Long-run supply of exports
Swift (1998) and Dvornak et al. (2005) find that there is empirical support that Australia is not a price taker in global markets for non-commodity exports. Following this, we assume domestic firms have some market power, with export prices determined by the cost of inputs. Specifically, supply of noncommodity exports is modelled via a log linear price mark-up equation, where the price of good k is a mark-up over nominal unit labour costs (labour cost per unit of output) and intermediate goods costs per unit of output:
28)
where w is the wage rate for good k, n is the level of labour input for good k, y is the level of output, pi is intermediate price, and i is intermediate volume, 0<ϴ<1. This framework is further simplified by assuming that there is no substitution in intermediate goods, so i/y is fixed which suggests[5]:
29)
Assuming the prices of domestically produced intermediate inputs are governed by a similar cost markup equation it follows that the export price can be written in terms of nominal unit labour costs and imported intermediate input prices:
30)