Initial Conditions and Post-entry Performance:

the Case of Indian Software Industry

Surendrakumar Bagde

Heinz School, CMU, Pittsburgh

sbagde@andrew.cmu.edu

In this paper we study entrepreneurship and firm survival, using as our empirical backdrop the Indian software service industry.  Four types of firms comprise the set of software exporting firms: foreign firms using India as a platform for software development (branches), firms founded by the Indian diaspora, typically based in America (diaspora), existing Indian firms diversifying into software production, from related and unrelated markets (diversifiers) and de novo startups (startups and spinoffs, as appropriate). Foreign firms enjoy superior access to the export market, established firms can count upon their existing managerial and marketing capabilities, and Indian firms have a greater familiarity with local labor markets and relevant local conditions.  Disentangling their impact on survival is challenging, not the least because of scale effects.  It is well known that larger firms are less likely to exit, especially in the software services exports sector, where there are many advantages of scale (cf. Arora and Asundi, 1999).  Insofar as certain types of firms are also more likely to enter at a larger scale, it is important to empirically distinguish between the effects of the scale of entry and the firm type. For instance, Roberts et al. find that firms that are more efficient are also more likely to choose a large scale of entry.

We develop a simple model, with forward looking firms that invest in marketing, and where current sales condition the distribution of future sales (thereby building in a direct advantage to large scale entry).  Our model nests several theories of firm entry and growth that have contrasting predictions about the impact of the scale of entry on firm performance, after conditioning on lagged output.  If efficient firms choose larger output at the time of entry, then it follows that, in a model where current sales also condition future sales, both initial and current output should reduce exit. On the other hand, a model in which initial output is exogenously given (for instance, by the size of the initial export contract the firm was able to get), the prediction is the opposite: More efficient firms would be willing to enter even with a small scale of initial output.  We test these contrasting predictions with a hand collected data set of software exporters registered with the Software Technology Parks (STP) NOIDA, part of the second largest software export cluster in India, between the year 1991 and 2003.

Though Indian software exports have grown steadily at over 35% per year over the last 15 years, our data indicate substantial exit as well.  We collected data on export revenues, information on background of founders, and year of entry. We study survival of the firms using Cox proportional hazard regression. After controlling for the age and other characteristics, and conditioning on lagged output, we find that the entry size does not influence exit.  In addition, foreign firms and diaspora firms have lower probability of exit, even after accounting for the scale effects. However, when “fly by night” firms are included in the sample, initial scale reduces exit, even after conditioning on lagged scale.

Our findings have two important implications. First, it appears that software exporters in India were very uncertain about their efficiency, especially in the early years of the industry.  Second, access to the markets is more important than local knowledge, in the software export industry. However, since diaspora firms perform better than foreign firms, local knowledge is also beneficial.

References:

Arora, Ashish and Jai Asundi (1999), "Quality Certification and the Economics of Contract Software Development A Study of the Indian Software Industry," NBER Working Papers 7260, National Bureau of Economic Research, Inc.

Roberts, P.W., Steven Klepper, and Scott Hayward, “Founder Backgrounds and the Evolution of Firm Size and Scope” Preliminary Draft, Carnegie Mellon University, Pittsburgh.