8. Develop a global partnership for development
9.4 Is the focus on SME fi nancing in developing
We have seen that considerable effort and resources have gone into addressing the financing gap for SMEs in developing countries, and that these efforts are justified on the basis of the unique contribution that the SME sector is sup- posed to make to the economy. In particular, the consensus has been that:
1 SMEs are responsible for the bulk of new job creation. Stimulating the sector can therefore have a disproportionate impact on employment and poverty levels.
2 SMEs are a fundamental source of entrepreneurship and innovation, and a vibrant SME sector is therefore likely to result in higher future growth rates.
Table 9.1 Proportion of banking assets owned by state, 2003*
India** 75%
Togo 51%
Tunisia 43%
Germany** 42%
Russia 36%
Liechtenstein 32%
Argentina 32%
Taiwan 28%
Poland 24%
Mali 22%
Source: World Bank Financial Structure Database.
* State has more than 50% stake.
** As of end 2002.
3 A well-developed SME sector increases levels of competition in the economy, by raising the degree of dynamism and flexibility therein.
Although this has long been the received wisdom, some relatively recent research has begun to question these assumptions. Biggs (2002) sets out to test the evidence on each of these claims, with interesting results.
9.4.1 Do SMEs create the bulk of new jobs?
Much of the evidence on SMEs’ role in job creation comes from studies in the US. In particular, the work of Birch (1979, 1981, 1987) is the key source in this regard. Birch’s research into the US economy in the 1970s produced the eye-catching finding that 80% of all new jobs in the US during this decade were created by companies with less than 100 employees.
However, the validity of this finding has been questioned on a number of grounds.
First, Birch’s classification of SMEs included small establishments started by larger companies. For example, the opening of a new Wal-Mart with around eighty employees would be classified as job creation by SMEs.23
Second, Birch’s analysis focused on gross job creation, rather than net job creation. It is widely accepted that SMEs do create a large proportion of jobs;
however, due to the high failure rate in the SME sector, they also destroy a large number of jobs. When net job creation is considered, the evidence suggests that large companies create more permanent, stable jobs than does the SME sector24 (Biggs, 2002).
Although the majority of the data on this issue pertains to the US mar- ket, the limited number of developing country evidence also supports this conclusion. Biggs and Shah (1998) examined the experience of four sub- Saharan countries in the 1980s and found that large enterprises (>500 employees) were the dominant source of net job creation. The precise figures were 56% in Ghana, 74% in Kenya, 76% in Zimbabwe and 66% in Tanzania.
As well as the quantity of net jobs, large firms are also responsible for the highest-quality job creation. In both developed and developing countries, large firms pay significantly higher wages (35% premium in developed coun- tries and up to 50% in developing countries), and offer better benefits such as pensions, as well as providing superior working conditions (ibid.).
9.4.2 Are SMEs ‘seedbeds’ for entrepreneurship and innovation?
First, do SMEs disproportionately contribute to innovation within the economy? For Biggs (2002) it entirely depends on the sector. That is, SMEs are indeed more innovative in the hi-tech, skill-intensive sectors such as computing, but large enterprises produce the majority of innovations in lower-technology, capital-intensive sectors such as food and chemicals.
However, it is likely that even in the narrow hi-tech sector, these findings from the US are not transferable to developing countries: ‘At earlier stages of development technology transfer from abroad is the force driving techno- logical progress, not innovation. Very little investment in basic R&D is undertaken by either large or small firms in low-income countries.’25 That is, the focus in most developing countries is on acquiring and adapting foreign technology. The example of the East Asian economies, starting with Japan and later incorporating Singapore, Taiwan and South Korea in particular, is instructive in the regard.26
Figure 9.4 illustrates this discrepancy in R&D spending, as seen from the late 1990s. As we can see, spending on R&D in high-income countries averages a little under 2.5% of GDP, compared to a little over 0.5% in middle- and low-income countries.
Within the developing world, the East Asian region had the highest relative research expenditure (0.9%), followed by Eastern Europe (0.88%), South Asia (0.75%) and Latin America (0.56%). The World Bank holds no data on North or sub-Saharan Africa in this area, but it is safe to assume that the level of R&D expenditure must be lower than the other regions for which data is available, as presumably this is what pulls down the low- and middle-income averages to the levels shown above.
One of the principal means through which technology transfer can be achieved is via linkages with international TNCs operating in the coun- try concerned. Indeed, research in sub-Saharan Africa,27 Asia and Latin America28 suggest that it is these linkages that give larger firms a productivity advantage over SMEs in these regions. It is therefore difficult to see SMEs themselves as ‘seedbeds’ of innovation in the developing world. Furthermore, to the extent that SMEs are involved in the export sector, this is in an indirect
Figure 9.4 R&D expenditure as percentage of GDP, 1998–2000.
Source: World Bank WDI.
form through their role in the supply chain of larger companies, which are often also TNCs (ibid.).
The argument that SMEs are responsible for the majority of new firm formation has already been considered in the context of job creation, but a similar argument holds for firms’ survival and longer-term growth prospects.
That is, SMEs create lots of new firms, but a high proportion of these fail and/or are unable to grow into medium-sized enterprises. For example, fewer than 10% of African firms with fewer than ten employees ever grow to employ more than fifty people (Biggs, 1999). In part, these low survival and growth rates may be because the establishment of a small enterprise is often a last resort. That is, the lack of alternative forms of employment may drive the desperate to set up on their own, which is supported by evidence from both developed and developing countries that the level of SME start-ups increases in economic downturns.29
9.4.3 SME dynamism vs. large companies economies of scale
It is certainly the case that SMEs add to the dynamism and flexibility of an economy, particularly in hi-tech sectors. Historically, this has led many to conclude that SMEs therefore also contribute to higher levels of growth – a subject we shall return to below. However, as well as this dynamism and flexibility, an economy also needs an appropriate quantity of larger firms that are able to benefit from the exploitation of economies of scale.30 There is thus a balance to be struck. In developing countries, on average, 60% of economic activity is generated by the SME sector. For developed countries, the corres- ponding figure is 30%. This suggests that, if anything, developing countries need to increase the proportion of larger firms within their economies.
Figure 9.5 Proportion of total employment by SMEs in selected developed econ- omies, 2001.
Source: Biggs, 2002.
However, as shown in Figure 9.5, even within developed countries there are wide differences in the size of the SME sector, which this average figure disguises. For example, in the US small firms have an employment share of 35%, which is the same level as in Japan. Sweden, in contrast, has 25% of jobs in the SME sector, whilst Germany has 57% and Italy has 61%.
At the other extreme, in Eastern Europe in the early 1990s as few as 1% of jobs were in the SME sector, reflecting the legacy of the communist era, when the focus was on large-scale, centralised production and the exploitation of economies of scale. Conversely, in many developing countries, the problem has been too little focus on large-scale production, which may be an after-effect of widespread import-substitution-industrialisation, where the monopoly pos- ition of a few large but inefficient firms was protected, preventing other firms reaching a large size. At the same time, excessive regulation and high tax levels encouraged many small firms to remain in the informal sector (ibid.).
9.4.4 SMEs and growth
The findings presented above – largely from Biggs (2002) – suggest that the focus on SME development may not be quite as justified as has been thought.
However, much of the evidence is somewhat anecdotal and/or ad hoc, and cannot be said to be conclusive evidence against the importance of the SME sector in the aggregate sense.
In an attempt to address this, the World Bank developed an internationally comparable database of the size of SME sectors in a range of developed and developed countries, and sought to develop statistically robust evidence on their impact upon employment, growth and poverty reduction.31 The study finds that:
• Contrary to the data provided in Biggs (op. cit.), the size of the SME sector – in terms of contributions to employment and GDP – shows no variability across countries at different income levels. However, this is only so if the informal sector is included, which was not the case in previous studies. If one strips out the informal sector the findings – which again run counter to those presented above – show that the relative importance of the SME sector increases in line with per capita incomes.
As countries develop, however, the (informal) SME sector declines and the (formal) SME sector increases (Ayyagari et al., 2003).
• While there is a statistically significant correlation between the size of the SME sector and economic development and growth, causality tests sug- gest that it is not that a strong SME sector causes economic growth, but rather than as a country develops this tends to lead to an increase in activity in the SME sector (Beck et al. 2004b).
• Furthermore:
When the analysis focuses on income growth among the lowest
income quintile rather than the overall population, it again finds no evidence for the importance of SMEs. Nor does it find any statisti- cally significant relationship between the importance of SMEs and the depth and breadth of poverty across countries.
(ibid.: 3) Figure 9.6 illustrates this variable ‘relationship’, by comparing GDP per capita and the relative size of the SME sector in terms of employment in the manufacturing sector. As we can see, the countries considered do not, as the conventional wisdom might have predicted, exhibit any clear, positive rela- tionship between these two variables. Furthermore, as pointed out above, causality tests suggest that to the extent that any relationship exists between the size of the SME sector and growth, this runs from growth to SMEs, and not in the other direction – again, this is contrary to what the conventional wisdom would predict.
To summarise, the case for the unique benefits that SMEs can bring to the economy would appear to be overstated, particularly in developing countries.
However, it is clearly important for an economy to develop a vibrant and diverse economy. That is, a combination of competitive and efficient small, medium and large firms must be optimal. In this regard the real problem facing developing countries is that of the ‘missing middle’ described above, not necessarily the lack of SMEs per se. In this regard, SMEs do appear to face particular issues in growing their firm size and market share.
For Biggs (2002) an important obstacle to this process is ‘institutional failure’:
Figure 9.6 GDP per capita vs. SME manufacturing labour force share.
Source: Beck, Demirgỹỗ-Kunt and Levine, 2004.
The failure of public institutions in many developing countries to pro- perly enforce business contracts and property rights, and to provide adequate information on markets, raises the costs of governing market exchanges, sometimes prohibitively. In such high transaction-cost environments, the extent of the market for individual business firms is limited by the number of business transactions that can be governed by relational contracts. Firms often cannot take advantage of profitable opportunities outside their local networks of personal relations because of information and enforcement problems.32
However, it should be remembered that, as we have seen, these factors are no guarantor of economic success, nor is development necessarily dependent upon them in all instances.
However, it is surely the case that difficulties in obtaining finance are not the only constraints on the growth of small firms. The aggregated data does clearly show a relationship between the business environment and growth in this respect, which is what one would broadly expect: for firms to invest and grow they need to have confidence that (a) they can enforce their ownership rights, (b) contracts will be upheld, and (c) the business environ- ment in which they operate is relatively stable. Economies characterised by high levels of uncertainty and minimal enforcement of the rule of law are unlikely to facilitate this process.
Large firms are often able to surmount these problems by protecting their own property and investing in the acquisition of information. However, in most cases such expense is beyond the reach of the SME sector.
Empirical research in developing countries suggests that the economic suc- cess of SMEs is directly connected with the extent to which they have been able to overcome these institutional failures by becoming ‘embedded’ in these private institutional support systems developed by large firms (UNCTAD, 2000).
In this regard, the success or failure of an SME to prosper and expand may be strongly related to its ability to forge linkages with larger firms.
One such sector, of course, is foreign TNCs operating in developing coun- tries, particularly where TNCs are relatively large compared to the size of the national economy, as is the case in many least developed countries, for example.