High-growth firms: More reasons for caution?

Recently in this journal two brief Viewpoint articles have debated the increasing emphasis being placed on ‘high-growth firms’ in both the research literature and policy. The first was Derbyshire (2012) and the second was a response by Walburn (2012).

There was some agreement between the two. First, both agree that there is a gathering bandwagon building momentum behind high-growth firms as a policy panacea. Second, both Viewpoints urge caution in relation to shifting policy emphasis away from support for early-stage firms and onto high-growth firms as recommended by, for example, Shane (2009).

A third more implicit area of agreement is that both Viewpoints start from the premise that the concept of high-growth firms is a valid one. Derbyshire (2012) urged caution in terms of not shifting policy emphasis away from early-stage firms and onto high-growth firms, but this was not because the very concept of high-growth firms may itself be faulty. Rather, it was because there was argued to be a possible relationship between levels of competition among new firms and the subsequent emergence of high-growth firms. Similarly, Walburn’s (2012) Viewpoint also accepted a priori that the high-growth concept is valid and even went as far as to state that ‘There is general agreement about the importance of high growth small firms’.

This brief Viewpoint article takes a step back from this assumption by showing that, in fact, there is considerable disagreement in the literature about the importance and even the very concept of high-growth firms. Furthermore, this disagreement originates from some of the most prominent researchers in the field, including Storey (2011) and Coad (2009). Caution is therefore urged on policymakers embracing the emerging high-growth paradigm for reasons more fundamental than those described in either Derbyshire (2012) or Walburn (2012). If these prominent researchers are correct about the nature of firm growth then the very concept of identifying and supporting high-growth enterprises is faulty and perhaps even dangerous. If they are correct then switching policy support to high-growth firms would be tantamount to supporting firms with a lower propensity than the average firm to grow at a fast rate in the future.

The debate about firm growth has been usefully appraised in a recent paper by Storey (2011). It boils down to a question of whether firm growth is positively auto-correlated or instead subject to regression-to-the-mean or even negative auto-correlation.

Auto-correlation is a statistical property of a time-series of data in which a subsequent observation is related to previous observations. For a positively auto-correlated variable the next observation is likely to follow the same pattern as previous observations. So if a firm’s employment has grown for the last three years and firm growth is positively auto-correlated then the firm is more likely than not to grow in the next year too. If firm growth is positively auto-correlated in this way it is valid to identify and support high-growth enterprises because these enterprises have a greater probability than the average firm to grow at a fast rate in the future because they have grown at a fast rate in the past.

As Storey (2011: 306) points out, however, the evidence on auto-correlation is mixed at best. On balance, there is probably more evidence to suggest that firm growth is subject to regression-to-the-mean, or even negative auto-correlation. If firm growth is subject to regression-to-the-mean then a firm experiencing rapid growth over, say, a three-year period (the period over which high-growth firms are identified) is actually more likely to return to the industry average growth rate in the subsequent period than to go on growing at a rapid rate. As Storey (2011: 309) states by drawing on Parker et al. (2009): ‘…the most likely outcome following a period of fast growth, even over a four-year period, is that the firm will return to an industry norm’. Therefore, identifying and supporting high-growth enterprises would amount to identifying and supporting firms with a likelihood to grow at only an average rate in the future.

It could be worse than that however. Policies that identify and support high-growth firms could actually be identifying and supporting those firms with an increased probability of declining in the future. If firm growth is not subject to regression-to-the-mean, in which growth rates are pulled back to the industry average over time, but is subject to negative auto-correlation instead, then those firms growing rapidly over a three-year period actually have a greater probability than the average firm of declining in the subsequent period. As Storey comments (2011: 309) by drawing on Coad (2009), under these circumstances the most likely outcome is that growth in one period is associated with decline in the next.

Beyond those highlighted by Derbyshire (2012) and Walburn (2012) there are, then, more fundamental reasons for policymakers to be cautious about the gathering high-growth firms bandwagon. Policies targeting the identification and support of high-growth firms only make sense if firm growth is positively auto-correlated. In this case identifying fast-growing firms amounts to identifying those firms more likely to go on growing, and perhaps at a rapid rate, in the future. However, the evidence for positive auto-correlation of firm growth is patchy at best. On balance there is probably more evidence to suggest firm growth is subject to regression-to-the-mean or even negative auto-correlation. This would mean that policies designed to identify and support high-growth firms would focus support on those firms with a greater likelihood of growing at only an average rate, or even declining, in future.