Uganda ’ s agricultural export market choices : The interaction effect of export market information use

The study seeks to establish the interaction effect of export market information use on the relationship between international market selection and the choices of export markets for Uganda’s agricultural commodities. A cross sectional study format, heavily hinged on a hypothetic-deductive approach was adopted and the multiplicative effect was represented by Modgraph. The effect of systematic international market selection on export market selection differs as a function of export market information use. A positivist research paradigm guided the study and therefore a more qualitative perspective would compliment the results of this study. Secondly, multiple observations of the same variables over 5 to 10 years accounting for structural changes (that is, the creation of new trading blocks as well as new bi-lateral agreements) give a finer picture. Hosting sector specific international exhibitions and agricultural commodity exporters in Uganda would significantly affect export market choices and performance. This is the first study to test for the interaction effect of export market information use on the relationship between International market selection and the choice of export markets


INTRODUCTION
Empirical studies into International market selection have observed that firms select export markets by either adopting; a systematic approach, a non-systematic approach or a relational approach (Papadopoulos and Denis, 1988;Andersen and Buvik, 2002;Brouthers and Nakos, 2005;Papadopoulos and Martin-Martin, 2011;Musso and Francioni, 2012).
Similarly, several related studies have examined the factors that influence export market selection and choices (Papadopoulos and Denis, 1988;Sakarya et al., 2007), however, generalizability of their findings has been limited because they are not industry specific (Musso and Francioni, 2012) and according to Papadopoulos et al. (2002), -they haven't been tested sufficiently, and / or are too complex to apply in practice‖, and consequently, other sector specific studies have been sought for (Papadopoulos and Martin-Martin, 2011).In addition, existing studies have been silent on the extent to which Export Market Information-Use moderates the relationship between International market selection and export performance, despite the key role played by export market information in the selecting of export markets *Corresponding author.E-mail: ecelaaron@gmail.com.
Authors agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License (Vyas and Souchon, 2003;Williams, 2003;Toften and Olsen, 2003).In this regard, Williams (2003) points out -the competitive advantage resides in how the information is used, given that the same information is available to all competing firms‖.
Uganda's agricultural sector employs approximately 70% of the labour force (kasirye, 2011), and the sector is considered the backbone to the economy, contributing 37% of the nation's Gross Domestic Product-GDP (Price, 2013;Food and Agricultural Organization, 2013).The sector also affects the livelihood of almost 90% of the population directly or indirectly (Kraybill and Kidoido, 2009).As of 2012, Uganda earned $2.3 billion USD from its exports, representing a 423.2% increase in ten years, and it's worth noting that agricultural commodities have consistently contributed a significant portion of the country's export earnings.Given Uganda's heavy dependence on agricultural exports, ensuring sustainability and growth in agricultural export earnings would be in order.
Therefore, the choice of export markets is crucial for export performance.In fact, Papadopoulos and Denis (1998) revealed that export market choices determine success or failure in export activities.Similarly, Musso and Francioni (2012) consider International Market Selection (IMS) as the -most important decision in internationalization‖.In addition, the World Bank advises that its prudent to establish whether a -country has out or under-performed it's competitors in selecting high-growth destination markets and sectors‖.In addition, Juswanto and Mulyanti (2003) revealed that a country's export growth relative to the world average is heavily hinged on whether its -exports are going to relatively growing (stagnant) regions‖.
Therefore, whereas numerous lucrative export markets exist, most exporting firms can't be present in all these markets due to insufficient resources.Thus, the selection of appropriate export markets is essential for the success and survival of exporting firms (Douglas and Craig, 1992;Wood and Robertson, 2003).However, the selection process is no easy task for exporting firms (Papadopoulos and Martin-Martin, 2011).Despite the existence of a huge potential of trade in growing markets of Africa and Asia.The bulk of Uganda's trade is less diversified in favour of high growth markets.In fact, in assessing Africa's trade potential, Decreux and Spies (2012) revealed that, -Sub-Saharan Africa (SSA) needs to become less dependent on stagnating OECD markets (growing below world averages) of its traditional trading partners in the developed world‖.In this regard, studies like Brouthers and Nakos (2005), which advocate for the systematic approach, show that it is linked to better export performance, while other studies have revealed that insufficient export market information has increased the prevalence of the un-systematic approach to International market selection (Andersson et al., 2004).Advocates of the relational approach (Andersen and Buvik, 2002) observed that it is more likely to be applied in the context of manufacturing and service industries, which is contrary in the context of Uganda agricultural exports.They are predominantly in their primary form (un-processed and almost in their raw forms), for instance for Coffee, itscoffee, not roasted, not decaffeinated, for Cotton its cotton, carded or combed and for Tobacco its-unmanufactured, partly or wholly stemmed or stripped.Most notably, none of the previous studies considered the notion that relationship between International market selection and export performance can differ as a function of export market information use.In a bid to create an insight into Uganda's agricultural export markets, this study sought to; first, establish the relationship between Systematic, Un-systematic and Relational International approach to market selection affected export performance of Uganda's agricultural exporters.Secondly, to establish the moderating effect of export market information use in the relationship between systematic and export performance of Uganda's agricultural exporters.

The international market selection (IMS) concept
The decision about the ‗choice of export markets' is given significant weight in international marketing literature to the extent of being raised to the realm of strategic decisions (Papadopoulos and Denis, 1998;Brewer, 2001;Koch, 2001;Andersen and Buvik, 2002;Papadopoulos and Martin Martin, 2011;Musso and Francioni, 2012).Rahman (2003) attests to this by pointing out that; -there are more than 200 sovereign countries in the world… and no one firm is likely to have the resources to develop profitable business in all of them‖.Therefore, as Papadopoulos and Denis (1998) put it -its essential to develop an efficient and effective method for selecting a foreign target market."Wood and Robertson (2000) are in tandem by asserting -the question of how to select the right markets becomes increasingly important, in particular given the wide range of possible markets and the plethora of available information‖.To date, literature on International Market Selection has come up with various categorizations as to how International Markets are selected.The two traditional approaches; the Systematic and Un-systematic (Papadopoulos and Denis, 1998) and more recently, Andersen and Buvik (2002) forwarded a third approach, categorized as the Relational approach.

Systematic international market selection
The systematic approach to International Market Selection is guided by extensive analysis of potential markets in a structured and formalized method, and information used in this approach is about country and market factors.It involves a thorough examination of a number of factors before foreign markets are selected (Musso and Francioni, 2012;Jansen 2013;Andersen and Buvik, 2002).Papadopoulos and Martin-Martin (2011) describe the approach as one guided by -an ordered set of rules and procedures‖.
Similarly, Bouthers and Nakos (2005) refer to this approach as one guided by the use of an -objective criteria to select markets‖.The predominant factors considered for evaluation in this systematic and structured approach include: host country factors (market/ country attractiveness, competition in target markets, marketing infrastructure), firm specific factors (type of product, management characteristics, firm size and international experience) and entry barriers factors (country risk, tariff and non-tariff barriers and geographic distance).Andersen and Buvik (2002) make reference to the systematic approach, as one involving -statistical methods to analyse the potential target markets.Musso and Francioni (2012) referred to the systematic approach as one requiring ‗several analyses before selecting international markets'.H 1 : Systematic International Market selection positively affects export market choices

Un-systematic international market selection
On the other side of the continuum, lies the nonsystematic approach to foreign market selection.The impractically of the systematic approach lends credence to the non-systematic approach.Empirical evidence (Papadopoulos and Denis, 1988) has suggested that both small and large businesses use the unsystematic approach when selecting foreign markets (Papadopoulos and Denis, 1988).Musso and Francioni (2012) define the un-systematic methods as those guided by -non rational reasons that apparently defy the optimizing logic of the market‖.Factors that are routinely embed the unsystematic approach to international market selection include the choosing of markets on the basis of Psychic distance, serving markets as a result of unsolicited export orders or even deciding to export to a given market as a result of -word-of-mouth‖.Malhotra and Papadopoulos (2007) reveal the ‗wide-use' of non-systematic selection of foreign markets following a review of 31 empirical studies.
H 2 : Un-systematic International Market selection positively affects export market choices

Psychic distance
Perhaps no other construct has attracted much debate Ecel and Lindi 399 and inquiry in the area of international market selection than the ‗Psychic distance' construct.First brought into the realm on International marketing by Beckerman (1956), it has been a key ingredient of research into export market choices (Johanson and Wiedersheim-Paul, 1975;O'Grady and Lane, 1996;Stottinger and Schlegelmilch, 1998;Koch, 2001;Evans and Mavondo, 2002;Ellis, 2008, Ojala andTyrvainen, 2009).Central to the concept of psychic distance is the notion that firms select foreign markets based on similarities to their home markets.
Consequently, firms are likely to get into markets in their neighborhood, since -geographic proximity‖ implies greater access to information and knowledge (Andersen and Buvik, 2002).More recently, Sinha et al. (2015) assert that the role of psychic distance is increasingly witnessed in the choices of international markets.Similarly, Hakanson (2014) longitudinal analysis spanning 47 years, of the bi-lateral trade between 25 major trading nations attest to the presence of Psychic distance in the trade patterns exhibited, with the influence being more pronounced for goods that are standardized internationally.Perhaps, it explains why Angola imports most of its agricultural products from Brazil despite being geographical far apart (shared language, Portuguese which is hardly spoken in the rest of Africa).
However, several empirical studies have pointed to the insignificance of psychic distance in the selection of international markets (Dow, 2000;Mitra and Golder, 2002;Davidson and Brewer, 2001).For instance, Wood and Roberston (2000) revealed that exporters ranked cultural differences as least important when making export market choices.Sousa and Lengler (2009) study of Brazilian exporters also revealed that firms performed better in markets where they shared much less culturally.Such results highlight the inconsistencies in psychic distance studies and thereafter, its luck of universal generalizability especially in light of studies that posit the psychic distance paradox (Stottinger andSchlegelmilch, 1998: O'Grady andLane, 1996;Evans and Mavondo, 2002).It is therefore plausible that cultural differences will not affect the inclusion of such industrial buyers in the pool of potential export markets.It is against the earlier mentioned argument that the following hypothesis is made H 2 (a) Psychic distance hardly influences the choice of export markets

Unsolicited orders
Also known as -unexpected opportunities‖, unsolicited foreign orders came about as a result of a foreign customer initiating the export order (Ahmed et al., 2006;Beleska-Spasova and Glaister, 2011;Geishecker et al., 2012).Papadopoulos and Denis (1988) observed:

" … the choice of markets by first time exporters is often simply a reaction to stimulus provided by a change-agent.
According to this approach, an international market is chosen in a reactionary way or rather a passive way (liang, 1995;Lim, Sharkey and Kim, 1996;Pla-Barber and Escriba-Esteve, 2006), thus exporter is chosen rather than the other way round.Despite being associated with a reactionary attitude to exporting, unsolicited orders surprisingly (given that it is often associated with -low export involvement‖ (Dean et al., 1998;Williams, 2006)) account for a large proportion of export orders.For instance, while establishing whether recipients of unsolicited orders were randomly chosen, Liang (1995) established that unsolicited orders accounted for; -62 and 61% of exports in a Brazilian and Turkish study respectively, thus reaffirming earlier results of Bilkey (1978).A study of Australian exporting firms by Julain and Ali (2009), mentioned the strong link between a firm's -initial export involvement‖ and the -receipt of unsolicited foreign orders‖.The notion is in consonance with Aksoy and Kaynak (1994) study, although they further link the receipt of unsolicited orders to the -international orientation of the manager(s)‖.Muranda (1999) arrived at similar results in his study of Zimbabwean firms exporting textile and clothing and referred to unsolicited orders as a -motivating‖ factor in export operations.
H 2 (b) The number and frequency of unsolicited orders positively influences the choice of export markets

Relational international market selection
Whereas, the systematic and un-systematic approaches seek to find a country(s) to export to (unit of analysis), the relationship approach seeks to find an exchange partner with whom to establish a business relationship (Bradley, 1995;Andersen and Buvik, 2002;jansen, 2013).In this case, a firm searches for -feasible international exchange partners‖, with shared goals.Business allies, government business support programs, and previous customers makeup such partners.For instance, if a firm is benefiting from an export promotion program by a trustworthy partner like the government, it's likely to export to markets fronted by the government.The relational approach predicts that international export markets are selected on the basis of potential business partners or exchange / business partners with whom to do business.
Consequently, it is a common feature for national governments to support trade programs (Brewer, 2009;Freixanet, 2011;Wilkinson and Brouthers, 2000).From a national government's perspective, firms that are engaged in international business play a crucial role in the economic development of the nation, thus the need to provide assistance to such firms (Ahmed et al., 2002).Wilkinson and Brouthers (2000) revealed, -exporting firms are more likely to stay in business than nonexporting firms and achieve 20% faster employment growth‖.In this context, support from governments to these firms is manifested in the form of; the identification of target markets, evaluation of target markets, making firms aware of market developments and growth trends (Brewer, 2001;Ahmed et al., 2002).
Similarly, Wilkinson et al. (2000) noted that -firms are not islands', and therefore not self-sufficient‖.To a great extent, export promotion agencies act as trusted business partners to exporters, given the assistance rendered to enhance their performance.Jansen (2013) notes, "firms make country choices using knowledge and other resources of associated business unit, business associations, government agencies or other entities which they have shared interest".
Exhibitions and trade fairs are considered important sources of such business partners, given their tightly targeted and interested audiences (Sarmento et al., 2015).For instance, Kiryowa (2014) reported that the International Floriculture Trade Expo (IFTEX) that is routinely held in Kenya attracts buyers from over 40 countries.He further notes that one of the firms that participated (Rosebud-Flowers-Uganda) at the trade fair was able to increase their international sales by 10% after participating in the previous IFTEX's trade fair, and expanded their operations from 45 to 50 hectares.Similarly, Brewer (2001) noted that exhibitions were often used as informants especially when they needed to diversify into alternative markets.In their longitudinal study of in Portugal, Sarmento et al. (2015) noted that the exhibitions enabled participants to create business relationships, given the face-to-face interface with prospective buyers, that is, this aspect is in tandem with the gist of relational international market selection which seeks to create relationships with trusted business partners in their networks.It is against that background and arguments that the following hypothesis is made;

Export market information use as a moderator
Foreign markets are often unfamiliar territory and characterised by business uncertainty, therefore information about foreign markets is key to dealing with such uncertainty (Diamantopoulos et al., 2003;Vyas and Souchon, 2003).In fact Tesfom and Lutz (2006) cites -insufficient information about overseas markets‖ as a significant deterrent to exporting by many firms (small) in Sub-Saharan Africa.
Similarly, Wood and Robertson (2000) assert that information is necessary when choosing international markets.However, recent studies in the realm of Export Market Information have shifted from the mere possession of information about export markets (acquisition), to actually using the said information.This position is eloquently summarized by Diamantopoulos et al. (2003) who equate it to -firing a rifle blind‖, in which case hitting the target is a matter of luck.Similarly, Diamantopoulos and Souchon (1999) observed, "any information acquired by decision makers will bear little impact on ultimate company performance if it is not actually put to use in the making of decisions.As the same information may be available to competing companies at about the same time." Accordingly, export market information can be used varyingly.Under circumstances where it is used to get a solution to a specific problem, with intent of immediate use, the kind-of-use is termed as Instrumental-Use (Diamantopoulos and Souchon, 1999;Toften and Olsen, 2003;Williams 2003).Subsequently, the use that is not geared towards any particular issue or problem at hand, but rather for purposes of -general enlightenment‖ (Williams, 2005) or rather widening the -managerial knowledge base‖ or even -stored for future use‖ (Diamantopoulos andSouchon, 1999: Vyas andSouchon, 2003) is termed as Conceptual-Use and lastly, Symbolic-Use is consistent with distorting findings, non-use of information, distorting information to suit already made decisions (Toften and Olsen, 2003), using export market information to build confidence in decisions already taken and for self-promoting motives as well as haphazard use of export market information (Vyas and Souchon, 2003) or even the use of market information to give credibility to decisions made on the basis of instinct (Diamantopoulos and Souchon,1999).
Whereas, consensus exists with regard to distinction between the Symbolic-Use and the Instrumental use, differences exist with regard to Instrumental-use and/or the conceptual-use.Some scholars believe that they are separate entities whereas others consider them as the same (Diamantopoulos and Souchon, 1999).With guidance from Diamantopoulos and Souchon's (1999) study that aimed at the measurement, scale development and validation of the export-information-use construct, this study adopted two dimensions of export-market information use, that is, instrumental/conceptual-use and symbolic-use, this approach was also adopted by Toften and Olsen (2003).They held the view that -instrumental and conceptual uses were actually aspects of a single dimension‖.
A key feature of the systematic approach to international market selection is the prevalence of -extensive analysis‖, and for analysis to take place the presence of information often precedes.For example, firms applying the systematic approach look into information that deals with the firm's owner potential to exploit international opportunities, market potential (growth rate) or even the competition in the potential or target markets (host-country factors), and information that tells the prospective exporter about the barriers to international markets (tariffs and non-tariff barriers).The above facets of selecting markets systematically all point Ecel and Lindi 401 to the critical role of information and more specifically the instrumental/ conceptual use of the said information, that is, particular problem or issue for which information is needed is criteria for choosing an export or the need to evaluate potential foreign or international markets among a pool of possible markets worldwide.Although previous studies in International market selection have taken into account the need to have the above kind of information, they have been silent about the way such information is used and how -use‖ consequently affects the relationship between selection criteria and the export market choices made.It is against this argument that following hypothesis is made.
H 4(a): Export Market information-use moderates the relationship between systematic international market selection and the choice of export markets.

METHODOLOGY The research paradigm and design
A cross-sectional study format was adopted in this study and guided by a positivist research paradigm, heavily hinged upon a hypothetic-deductive approach (that is, a quantitative operationalization of concepts and the formulation of hypotheses that are statistically tested from a large sample).
Population, sample size and sampling procedure 563 firms exporting agricultural commodities formed the population of the study; this list was obtained from the Uganda Export Promotion Board's (UEPB) database.Often, the sample size for a given study is determined by either using formulas to ran a calculation and come up with a sample (Yamane, 1967) or by referring to published tables that set sample sizes at given levels of population sizes for example, in Krejcie and Morgan's (1970) formula for sample size determination (Israel, 1992;2013).
In this study, the researcher applied both approaches and eventually considered the approach that resulted into a higher / greater sample size.At a 95% confidence interval and variability of 0.05 level of precision, the Yamane (1967) formula resulted into a sample size of 233.8, whereas Krejcie and Morgan's (1970) table of sample determination informed that for a population (N) between 550 and 600, a sample (n) of 226 is representative of the population.
Consequently, the former approach was considered for this study.Commodities exported by the firms where grouped based on the harmonized commodity description and coding system (HS), which is an internationally standardized system of names and numbers to classify traded products, thus creating mutually exclusive, homogenous strata.This classification was done at the 2-digit chapter level of the HS coding System.Given that the contribution to the nations export earnings export earnings by the different export commodities is not proportionate, with exports like coffee and tea contributing more, a dis-proportionately stratified sampling technique was adopted.The exporting firm and their managers or executives directly involved in the export functions of the firm constituted the unit of analysis and unit of inquiry respectively.

Measurement and operationalization of the variables
The constructs under study are both theoretical and unobservable, and therefore necessitating an operationalization to develop indicators that measure the constructs.Multiple indicators were deemed appropriate given that the variables were in the social science domain, and therefore very susceptible to subjectivity and imprecision (Bhattacherjee, 2012).The items were anchored in a 5point likert scale (1= Strongly Disagree to 5 = Strongly Agree).

Systematic international market selection
The gist of systematic international market selection is objectivity and structure, coupled with a formal process (Andersen and Buvik, 2002;Brouthers and Nakos, 2005;Musso and Francioni, 2012).Accordingly, factors relating to systematic international markets selection in this study were conceptualised to comprise of; firm size, managerial international experience, market or country attractiveness, extend of competition in the foreign markets, the extent of tariff and non-tariff barriers, and the presence of preferential market access provisions (Brouthers and Nakos, 2005;Alaoui and Makrini, 2014;Koch, 2001, Lages et al., 2008;Del Rio Araujo and Neira, 2006;Moen, 1999, Cavusgil andZou,1994).

Un-systematic international market selection
The non-systematic approach to international market selection is characterised by a personalised and irrational approach to the selection of international markets.Alexander et al. (2007) refer to this approach as a -subjective evaluation of possible alternatives‖.Andersen and Buvik (2002) refer to the approach as one that involves -expanding internationally on an opportunistic basis‖.Accordingly, psychic distance, unsolicited orders and word-ofmouth were the dimension used to captured aspects related to unsystematic International market selection (Brewer, 2001;Liang, 1995;Muranda, 1999;Gieshecker et al., 2012;Papadopoulos and Denis, 1988;Leonidou et al., 2007).

Relational international markets selection
Measurement of relational international marketing was guided by Andersen and Buvik (2002) study, which introduced the notion of relationships with trusted business partners as means of selecting foreign markets.Wilkinson et al. (2000) allude to role of -other firms and organisations‖, and note that governments have become increasingly involved in boosting the international performance of their firms.Accordingly, relational international market selection was measured in terms of; government business support programs, exhibitions/trade fairs and previous customers.Items were adopted form others studies like Ellis (2000).

Export market information use (Moderating variable)
Export market information use is operationalized under the premise that the same marketing information is accessible to all competing firms at the same time, thus the competitive edge is based on usage rather then acquisition (Diamantopoulos et al., 2003).The -use‖ of export market information has been conceptualised as either Instrumental/ conceptual or symbolic (Diamantopoulos and Souchon, 1999).Items scaled were adopted and adapted from related previous studies including; Toften and Olsen (2003), Vyas and Souchon (2003), Williams (2003) and Toften and Rustard (2004).

Export market choice
The gist of the study involves establishing those factors that influence the export market choices made by agricultural commodity exporters in Uganda.In order to determine how various factors affected the choices of export markets made and given the heterogeneous nature of foreign markets, indicators that created a distinction in the markets chosen were developed mainly in line with export performance literature.Choice implies the presence of a variety, and therefore necessitating a selection.In particular; from an exporting firm's perspective; growth in export sales, growth in market share, business expansion to meet export demand and profitability (over 5years) were considered sufficient distinction for the export markets.Export market choices made affect export performance (He and Wei, 2013) therefore adopting an export performance distinction of the choices made.When a particular export market is growing / promising, chances are that export orders will rise thus the growth in a firms export sales and the need for business expansion to cater for the growing demand.However, stagnating export markets will imply reduction in export orders, which will consequently affect export sales, market share and expansion plans and profitability.

Data collection and management
In accordance with the research paradigm adopted, (Positivism), quantitative data was collected from the respondents by means of self-administered questionnaires.This exercise was carried out between June to August 2015.To minimize the effects of bias resulting from the -form of measurement‖ used (Common Method Bias or variance), procedural remedies were applied in accordance with Podsakoff et al. (2003).Whenever possible, responses dealing with the export market choices were obtained from another respondent in the export department, in order to reduce consistency motif.However, time lags were frequently used to create temporal separation after responses on the International Market Selection approaches were gathered (brief conversations were struck up to achieve a time lag).
Additionally, assurance of confidentiality was repeatedly echoed to the respondents (Chang et al., 2010).Similarly, when using regression techniques in study, multicollinearity often affects the desired structural relationships.Paul (2006) notes that multicollinearity manifests when there is an exact or nearly exact linear relationship among two or more of the input variables.It is therefore prudent to establish the degree of multicollinearity.An examination of the correlation matrix to identity large correlation coefficients between predictor variables is simple method of detecting multicollinearity, and correlation coefficients next to unity are indicators of multiple collinearity (Farrar and Glauber, 2001;Field, 2009).
In this study, the regression results were valid given that the correlation coefficients between the predictor / independent variables were below 0.8 in tandem with Field (2009).For data reduction, a factor analysis was carried out, and the results in Table A1 show the factor analysis of international market selection and indicate that 8 items loaded on un-systematic international market selection, while 7 items loaded on systematic international market selection and 5 items loaded onto relational international market selection.In total, international market selection was examined by 15 scale items that accounted for 63.3% of the variance in international market.The results in Table A2 show the factor analysis of the variable export market information use.The results indicate that 4 items loaded on instrumental use and 3 items loaded on symbolic use and accounted for 61.4% of the variance in overall export market information.Lastly, the results in Table A3 show that 5 items examined the choice of export market and accounting for 62.2% of the total variance in the overall choice of export market.

Sample characteristics
Of the 234 commodity exporters targeted, 44 questionnaires were incomplete and therefore 190 responses were fit for the analysis, consequently obtaining a response rate of 81.1%.The majority of the respondents were male (70%), and worked at the middle level of management (54.2%).The majority of the respondents (46.3%) were between the group of 36 to 55, while 45.7% of the respondents had attained a university degree.70% of the firms had been in existence for over 10 years, while a greater proportion (73.6%) of the firms exported to destinations where they had preferential market access.Most of the exports (44.7%) were destined for the East African Community (EAC) market that is, Kenya, Tanzania, Rwanda and Burundi, with the European union representing 15.8% of the export destinations.

Correlation analysis results
Zero-order correlations (Pearson's) were used to establish association between systematic, unsystematic and relational international market selection and the choices of export markets.The results presented in Table 1 indicate a positive and significant relationship between systematic international market selection and the resultant export market choices made (r = 0.298**, p < 0.01).Thus implying that firm specific, host country and entry barrier elements positively influence export market choices of agricultural commodity exporters Uganda.This is consistent with those of similar studies.For instance, Brouthers and Nakos (2005) observed that, systematic international market selection significantly influenced export performance.Similarly, Ahmadian and Ma'atoofi (2011) concluded that for Iranian exporters, the -willingness and conviction‖ to enter various export markets differed as a function of the managers' International experience.
However, contradiction with the findings is found in the works of Pla-barber and Alegre (2007).They focused on 121 exporting firms in France and established that firm size hardly influenced export market choices.Similarly, Calof (1994) observed, -due to modest nature of the variance explained by firm size in export behaviour, it's importance is inconsequential‖.
Upon examining the relationship between un-systematic international market selection and the export market choices made, the results in Table 1 revealed a negative and significant relationship (r = -0.232**,p < 0.01).Implying that when irrational factors guided the export market selection, a detrimental effect was observed in the choices made.With regard to the relationship between psychic distance and the choice of export markets, the results show an insignificant correlation (r= 0.139, p > Ecel and Lindi 403 0.05) and therefore supporting hypothesis H 2(a) that states that ‗Psychic distance hardly influences the choice of export markets'.Therefore, despite the popularity of the of the psychic distance concept in international market selection (Berkerman, 1956;Johanson and Wiedersheim, 1975;O'Grady and lane, 1996), such contradictory results are not entirely new.For instance, Brewer (2001) arrived at the conclusion that psychic distance did not play an important role in the firms' international market selection decisions.The results in Table 1 also indicate that unsolicited orders negatively and significantly influence the choice of export markets (r = -0.204*,p >0.01), consequently not supporting hypothesis H 2b , which states -The number and frequency of unsolicited orders positively influences the choice of export markets‖.This implies that the reliance on unsolicited orders by Ugandan agricultural commodity exporters negatively affected the choice of export markets made.
This finding is in contradiction to Geishecker et al. ( 2012) and Leonidou et al. (2007) findings who attest to the popularity of this approach as one of the most influential factors driving international sales for most exporting firms.Plausible explanations to the contradiction in results include; Bilkey (1976) who observed that, because such firms (ones that rely on unsolicited orders) often take a reaction approach to exporting and just wait for -export opportunities that come their way‖, it is plausible that importers only resort to unsolicited orders when their main (previous) suppliers have had an interruption in supply, and resort to unsolicited orders as a temporary measure.Secondly, Liang (1995) observed that the frequency of unsolicited orders depended a lot on the uniqueness of the product.Therefore, it is conceivable that the unprocessed nature of the agricultural commodities exported by Uganda, reduces their chances of uniqueness.
The results in Table 1 also show a positive and significant relationship between relational international market selection and the choice of export markets (r = 0.255**, p > 0.01), This implying that the use of government support programs, exhibitions and trade fairs as well as previous customers is strongly associated with favourable export markets.Thus, supporting hypotheses H 3 .The findings are consistent with those of Bradley (1995) and Andersen & Buvik (2002) who emphasize the role of partners in the selection of export markets.
Similarly, in the specific context of Uganda's flower exporters, Kiryowa (2014) revealed that participating at exhibitions significantly marketed Uganda flower exporters with up to 20% growth in direct sales.Additional support found in Rose (2007) study observed a positive link between the number of embassies or consulates and the growth of a nations exports.

Moderation (Interaction) results
The results in Table 2 show that systematic international   1 indicate that the effect of systematic international market selection to the choice of export markets is different at various levels of export market information use, given the differences in gradients of the graphs.Accordingly, since the magnitude of the effect is greater at one level compared to another level, it is plausible that the interaction effect is significant, and thus supporting hypothesis H 4(a).The interaction term was antagonistic in nature and therefore its presence reversed the effect of systematic IMS on the choice of export markets.These results are supported by previous studies; for instance, Vyas and Souchon (2003) noted that market information use was especially vital in the export market setting given its -diversity and turbulence‖.Diamantapolous and Souchon (1999) allude to the fact that instrumental use of information about exports markets result into better choices in such markets and the subsequent competence of the exporting firms.Toften (2005) also concurs with this point of view.

Conclusion
The findings of the study revealed that export-market- information-use only moderates the relationship between systematic international market selection and the choice of export markets, and had an insignificant interaction term for Relational International market selection and the choice of export markets.The results also observed a positive and significant relationship between systematic international market selection and export market choices.Additionally, the findings revealed that the psychic distance effect was inconsequential to export market choices, while unsolicited export market orders had a significant and negative influence on export market choices.A positive and significant relationship was observed between relational international market selection and export market choices.

Theoretical implications
Theoretically, the study shows that interacting systematic international market selection and export market information use provides a significant explanatory power to export market choices.Previous empirical studies into international market selection acknowledge the crucial role of information in export market selection, but such studies fell short of establishing if the way such information is used (export market information use) had an effect on the relationship between approach and the resultant choice.As a modest contribution to theory development, this study borrows from assertions by Williams (2003) and Vyas and Souchon (2003) that, use rather than possession of export market information provides the competitive advantage.This addition is a substantially significant departure from previous studies in the area of international market selection.
Secondly, the psychic distance construct is an instrumental component of international market selection research, and it attributes export market choices to the perceptual differences between the home market and a foreign market business and cultural practices.However, this study has presented empirical evidence to support the notion that the psychic distance concept is "past its due date" as posited by O'Grady and Lane (1996), Stottinger and Schlegelmilch (1998) and Evans and Mavondo (2002).In so doing, affirming that the psychic distance effect is inconsequential when applied to international market choices, specifically in the context of agricultural commodity exports from a developing country context.This is in tandem with Brewer (2001) who established that psychic distance hardly affects export market choices of Australian exporting firms.

Policy and managerial implications
First, the study revealed that government support agencies significantly influence export market choices.Although the ministry of trade and the Uganda export promotion board thrives to promote exporters, the country's trade facilitation structures are still at infancy to be able to support all exporting firms to their full potential.Therefore, consulates and missions abroad ought to actively engage in trade facilitation in their various potential export markets by especially reducing information asymmetry.
Secondly, the agricultural exports sector ought to host an annual international trade exhibition targeting international buyers.This strategy has already been adopted by the Kenyan flower exporters through the hosting of the Annual International Floricultural Trade Expo (IFTEX) that attracts buyers from over 40 countries.By hosting sector specific exhibitions of international magnitude, small and medium sized exporters without sufficient logistical support to attend exhibitions outside the country, would get exposure to other buyers and possibilities for export market diversification.
The findings also observed that export market choices guided by unsolicited export orders were unsustainable, therefore export managers ought to consider such unsolicited orders as complimentary rather than as a major or predominant method in international market selection.

Study limitations and possible areas for further research
The study is not devoid of shortcomings and therefore limits generalizations.For instance, results from a qualitative research design would compliment those of this study that was guided by a positivist research orientation.Secondly, given the cross-sectional nature of the study, repeated or multiple observations of the same study over a longer period (5 to 10 years) would account for structural changes in the environment that the respondents operate in (i.e.bi-lateral trade agreements).Lastly, whereas this study established a positive link between government support programs and export market choices, further research ought to be done to explore methods/ modalities of integrating strategic position of embassies and high commissions into the export promotion cause or export market penetration endeavours of exporters.

Variable
Market growth Export market share We have expanded our operations in order to meet demand in most of export markets 0.646 -We have been profitable in most of our export markets in the last 5 years 0.798 -Over the past 5 years, we have increased the number of export markets we serve 0.744 -Our export sales have grown in most of our export markets over the last 5 years -0.752Our Market share has improved in most of our export markets over the last 5 years -0

Figure 1 .
Figure 1.Interaction effects of systematic IMS and Export Market Information Use on the selection of export markets

Table 1 .
Zero order correlation between International market selection and export market choices

Table 2 .
Hierarchical regression of Interaction effects of Systematic International Market Selection (IMS) and export market information use on the choice of export markets

Table 3 .
Factor analysis for choice of export market.