Software export success factors and strategies in ‘follower’ nations

Abstract

Many countries world-wide are becoming active or interested in software exports. This paper initially analyses the software export experiences of India, Ireland and Israel; the three largest software ‘followers’ (those which developed a software industry after the 1970s). It develops a new ‘Software Export Success Model’ on the basis of that analysis. This model, having been developed as a descriptive/analytical framework, is then applied in a number of ways. First, as a comparative framework, helping to identify current strengths and weaknesses of two ‘second-tier’ followers – Russia and China – as compared particularly with India. Secondly, as a predictive foundation for understanding likely sectoral trajectories within these second-tier nations. Finally, as a prescriptive foundation from which guidance can be derived on the strategies and interventions that these countries may need to adopt for software export growth. As well as demonstrating the value of this new framework for conceptual analysis, this paper also shows how it can have practical value to policy-makers, industry association representatives, consultants and others involved with software sector strategy. Although beyond the scope of this paper, it appears that the model will also serve as a framework for equivalent analysis and guidance for the many other follower nations who have recently entered into software exports.

Keywords:

• Software industry
• export strategy
• competitive advantage
• success factors
• India
• Ireland
• Israel
• Russia
• China

Acknowledgements

The authors are very grateful to two anonymous referees for their comments that led to significant revisions of an earlier draft of this paper.

Notes

Or, more accurately, few benefits that have been charted, since linkages between IT exports and local poverty alleviation remain largely unresearched.

Details of follower nation interviews and research project periods are: 1988–1989 (India – 65 software firm managers, 23 programmers, 30 government officials, 17 sector analysts, three association representatives); 1995–1999 (India – ten software firm managers, five programmers); 1998–2000 (India – ten software firm managers, four programmers); 2002 (India – four firm managers, one government official; Russia – one consortium representative; China – one association representative); 2003 (India – two software firm managers). Details of client/investor nation interviews are: 1988–1989 (UK – seven managers, three association representatives); 1995–1999 (UK – 15 managers); 1998–2000 (UK – seven managers); 2000–2003 (UK – 13 client firm managers, two consultants; USA – five client firm managers, three consultants; Norway – one client firm manager). Other findings from this primary research are reported elsewhere, e.g. Heeks (1996), Heeks et al. (2001), Nicholson and Sahay (2001), Lai et al. (2003) and Sahay et al. (2003).

Indian figures are for the financial year to end March (e.g. 2003 figure is for period April 2003 to March 2004). These figures represent gross earnings. Calculations based on earlier interviews suggest net earnings for India to be about 40 percent of gross because of outflows of foreign exchange from India to pay for: travel and living allowances of Indian software workers who undertake their contracts overseas, marketing, multinational profit repatriation, importation of hardware and software, etc. (Heeks 1999a). This estimate is not far off the more recent calculation by McKinsey which estimated that 67 percent of all offshore IT outsourcing expenditure from US firms is directly recaptured by the US economy (Olive 2003).

At least according to some figures (e.g. Woods 2004), the quoted levels of software export would mean Ireland is a larger software exporter than the USA (where, for example, official forms recorded exports of $US13 billion in 2002). Yet, the Irish software sector employs around 30,000 whereas the US software sector employs more than one million. Even allowing for a lot of Irish talent, luck and hard work, these figures just do not add up. Their explanation – both US and other leading economy under-reporting and Irish over-reporting – is transfer pricing by software multinationals booking exports in Ireland (which has low corporate tax rates) for products that were developed in the USA and other G7 nations. The true level of Irish software exports is hard to estimate. One approach would be to extrapolate figures from the indigenously-owned part of the industry (which is assumed not to transfer price). Assuming equivalent per employee export rates between the domestic- and foreign-owned parts (a questionable assumption, but better than ignoring transfer pricing), this would suggest exports in 2003 around the $US3 billion mark. This would amend the figures in Table 2 to 2.01 percent of GDP and 2.5 percent of exports. Other sources have halved the headline figure; for example, citing true exports in 1999 as c.$US4 billion (Kelly 2000; OECD 2000).

Interview data suggest variation depends partly on whether or not figures include software activity in non-software firms; partly on whether or not they include non-software staff working in software firms.

Development of the model has also benefited from feedback during its use in teaching, in conference presentation (Heeks & Nicholson 2002) and in field application (e.g. Nicholson & Sahay 2003).

In general, that export orientation has strengthened over time. In 1991, for example, the equivalent figures for exports as a proportion of total software revenue were: India c. 60 percent; Ireland overall: 93 percent; Ireland local-owned only: 41 percent; Israel: c. 20 percent (Heeks 1996; Arora & Gambardella 2004; NSD 2004).

Although, in India, examples of strong and sustained relationships between foreign multinationals and local software firms that did not involve equity investments were also found.

One might also wish to question the differentiation of the 3Is' profile into the stereotypical association of India with programming services, Ireland with package localization, and Israel with niche products. Software package exports form only a small proportion of the Indian total but still grew from roughly $US10 million in 1993/4 to nearly $US250 million in 2003/4 (Heeks 1996; Ahmad 2004). Investments in software R&D centres in India are also growing. Conversely, foreign investments in Israel have meant that much of its software work is actually equivalent to outsourced R&D services that are incorporated into foreign-owned packages (Teubal 2001). Finally, strip out the transfer pricing from Ireland's export figures and one sees in 2004 a much more diverse industrial profile that mixes elements of all three stereotypes. It is beyond the scope here to analyse this further but the evidence to date contains at least some hints of convergence.

For example, Ernst & Young (2003) cite average annual programmer costs in China of $US4,750 as compared to $US5,850 in India.

An open question remains in relation to US links. Silicon Valley's China-born population is second only to India-born (Kapur & McHale 2002); there are said to be more Chinese-origin heads of high-tech firms in Silicon Valley than India-born heads (Kharbanda & Suman 2002); and Saxenian (2002) reports higher return migration rates among Chinese than Indians. In theory, this should be fertile ground for the development of software exports to the US market but, in practice, it does not seem to have materialized to any major degree.

Calculated by dividing domestic software market size by population size. Actual figures are Russia ($US8.74 per capita in 2002); China ($US7.38 in 2003); India ($US1.88 in 2003/4).

Continuing the order of magnitude approach, one can define these as countries which make an export contribution of at least 0.005 percent of the global software market (equivalent to roughly $US20 million in 2003) but not more than the 0.05 percent achieved by second-tier nations. See Carmel (2003a) for an alternative taxonomy.