Are new organisations at the cutting edge of employment relations innovation?

Purpose – The purpose of this paper is to explore differences in the degree of innovation in employment relations (ER) between emerging and established firms, Design/methodology/approach – A large national telephone survey ( N ¼ 1,416) of both emerging ( o 5 years) and established firms was conducted. Findings – Emerging firms were more casualised, less unionised, and experiencing higher levels of market expansionand unpredictability. Despite these differences, younger firms showed otherwiseremarkable similarity to older firms across a range of ERpractices, and both categories showed a reliance on business networks, rather formal training, for ER knowledge. While introducing ER changes more rapidly than older (and larger) firms, they were converging towards a suite of ER practices similar to that adopted by older firms. The results suggest that, if anything, established firms may have been engaged in greater innovation in more unusual ER practices. Research limitations/implications – Only managers were surveyed. The data are cross-sectional rather than longitudinal. As the study was undertaken in only one country, replication in other settings would be desirable. Originality/value – The results raise major doubts about the notion that new firms represent the cutting edge of innovation, and highlights the degree to which newer firms match or mimic older firms ’ ER architecture.

they lack the stultifying institutional and cultural legacies of mature organisations. By that observation, it seems reasonable to expect them to be at the forefront of new methods of work and employment innovation. But are they? Within the broader innovation field there are persuasive arguments both ways (Katila and Shane, 2005). Within the employment relations (ER) literature there are some indications that younger (and smaller) firms do use innovative practices; and it is clear younger or smaller firms feature flatter hierarchies and more organic communication, thus creating a platform for innovative ER initiatives (Wilkinson et al., 2007b). The Australian Workplace Industrial Relations Survey (AWIRS) provides some empirical confirmationat least of differencesuggesting that new worksites have fewer occupational levels, fewer unions, and a lower incidence of industrial disputes (Morehead et al., 1997). Does this reduced union density associated with immature firms lead to the proliferation of alternative methods of organising and controlling labour and/or give rise to other forms of voice?
Alternatively, in the pursuit of competitive advantage, or reflecting exposure to similar ideas, it is possible that all companies end up converging on basically the same suite of ER measures (e.g Sparrow et al., 1994). While the convergence/divergence debate usually asks questions like "are organizations and managerial practices worldwide becoming more similar or maintaining their nationally based dissimilarity?" (Mcgaughey and De Cieri, 1999, p. 235), one could equally ask "do organisations converge towards similar managerial practices, regardless of age?" This paper describes the results of a national telephone survey of Australian business managers (N ¼ 1,416), and explores how the prevalence of particular ER features covaries with firm characteristics, in particular the youth of firms. This study controls for the issue of size both statistically, and by including only those new firms that have reached a quorum of 20 employees.
This study extends current research of ER in younger firms and compares responses from managers of firms of different ages, sizes and industry sectors. We will focus on new organisations rather than new workplaces per se. Our primary research question is: RQ1. Do new organisations show evidence of innovation and different practices when it comes to how firms manage their staff?
That is, are they at the "leading-edge" or alternatively, at the trailing edgesimply mimicking the behaviours of larger firms, to the extent that their more limited resources allow? Bacon et al. (1996) argue that owner-mangers are not merely picking-up new management practices as "flavours of the month". They suggest that "the new management agenda has penetrated deep into the UK economy and that innovative and progressive employee relations practices are no longer restricted to large mainstream companies" (Bacon et al., 1996, p. 87).

Innovation in ER
"Innovation" is a widely used but loosely defined term. It is often seen in an ER context to equate to practices that depart from traditional work and industrial relations norms. Much of the extant literature equates innovative practices with notions of flexibility and employee participation in an effort to create high performing workplaces (Ichniowski et al., 1996). An alternative, oft-cited definition by West and Farr (1990), suggests a product, process or idea only has to jump a low hurdle to qualify as an innovation: it only needs to be new to the "unit of adoption" (Becker and Matthews, 2008;Brav et al., 2009, p. 355). By this definition, new organisations might seem arithmetically certain to possess more innovation than old organisations in their ER practices, but such a conclusion would be banal and lead commentators to argue that rhetoric about innovation cannot "obscure fundamental continuities with the recent […] history of employment relations" (Giles et al., 1999, p. 15). Setting a higher bar, that innovations are not purely contextually 1398 PR 46,7 new, but have some intrinsic, broader novelty, as insisted upon in the OECD ("a new or significantly improved" change) (OECD and EUROSTAT, 2005, p. 47) does make the question of locating hotbeds of innovation more challenging, and opens the field up to broader debate. There are a priori reasons to expect more (or less) "higher bar" innovation in new organisations. Some scholars argue that, lacking the structures and procedures that might inhibit or domesticate novelty, new firms and workplaces are more likely to foster innovation (Henderson and Clark, 1990). The growth (and dynamism) (Kotey and Slade, 2005) associated with newness will at the least tend to disrupt ER, as Campagnolo and Camuffo (2009) argue: Employment relations become less sticky because changes in technology, production models and globalization systematically re-define the kind of knowledge, of experience and, hence, of workers, which is valuable for firms and markets (295).
Others suggest that new firmsif not new workplaceslack a body of existing knowledge or resources with which to build innovation (Cohen and Levinthal, 1990;Teece, 1992). In the field of ER practices, the difference between a "new" and an "old" organisation may disappear if one considers size as a confounding variable. Once a business grows to a certain size, new or old, ER may well have to be formalised, and professionals employed to supervise the increasingly complex and legally-constrained field of ER. The progenitor of business innovation theory, Josef Schumpeter, disparaged innovation that was reduced to "the business of trained specialists who turn out what is required and make it work in predictable way" (Schumpeter, 1942, p. 132). Yet that is one of the goals a professional ER expert strives for. The "routinization" of activities has been identified as an innovationsuppressant (Ohly et al., 2006), even when measured against objective external markers such as patent outcomes (Benner, 2009).
Innovation in ER may therefore equate in practice to the degree to which certain elements of formal ER are implemented in a given entity, either as independent measures or in "bundles" (Singh et al., 2012). Much of the focus in ER has been around the notion of high performance work systems but while much talked about these appear relatively rare in practice (e.g. Appelbaum et al., 2000;Lawler, 1992;Walton, 1980). Pfeffer (1998) lists seven "successful" practices: employment security, selectivity in recruitment, high levels of contingent reward, self-managed teams, extensive training and development, informationsharing, and harmonisation of status differentials. These are seen as being held together under an overarching philosophy with a long-term commitment and a willingness to engage in consistent measurement of whether or not high standards are being achieved. Other studies have different bundle configurations with wide variability in the number of practices included (Wilkinson et al., 2016) There is a second, less-commonly celebrated field of innovation in ER, relating to the impact of product or physical work-space innovation on ER. Research, for example, points to the impactspositive and negativeof hot-desking (Millward et al., 2007), the internet and e-mail (Lengnick-Hall and Moritz, 2003), teleworking (Golden, 2007;Taskin and Devos, 2005), mobile phones (Lowry and Moskos, 2005) and open plan offices (Hedge, 1982;Lee and Brand, 2005) on managing employees. A new organisation may engender novelty in ER simply through the way its structures, less systematically entrenched than in mature organisations, shape work practices.
Finally, a third identifiable source of innovation lies in the way employees and employers interactthe proliferation in new forms of employment arrangements, including part-time employment, temporary or contingent employment, contract employment, compressed work weeks, and job sharing (e.g. Campagnolo and Camuffo, 2009;Rau and Hyland, 2006). This is innovation as "flexibility", which may be to the advantage of the employer or, less commonly, the employee. These new forms of ER may force changes in how HR departments handle their work.
Greenfields sites, in these second and third senses, offer a particularly promising domain in which to investigate innovative work practices, with a completely new physical environment often accompanied by explicitly "new" forms of work arrangements for staff (even if they are borrowed from existing paradigms), and a recognised pathway chosen by management to actively remake work practices (Baird, 2001;Leopold and Hallier, 1999;Townsend, 2004;Waring, 2001). Researchers have explored work practices (Wilkinson and Ackers, 1995), recruitment and selection (Hallier, 2001), culture (Hursthouse and Kolb, 2001), ageing (Leopold and Hallier, 1999), employee consultation (Townsend, 2005) and teams (Mallon and Kearney, 2001) in greenfields sites, but these case studies do not tell us about the generality of the new organisation experience. Perhaps they attract attention precisely because they are unusual in some way, rather than because they represent how new organisations operate. Debate continues around whether new firms, sites or greenfield agreements enable managers to organise work more flexibly, better reward employees, tailor pay and conditions to individual circumstances and motivate higher productivity and pay ("high road" approaches), or lead to the abandonment of structured systems of personnel management, a loss of employee conditions, and a drift of labour from small to large organisations "low road" (Gahan, 2007;Youndt et al., 1996).
In Australia, the registration of greenfields agreements with the Fair Work Commission and its predecessors provides one indicator of incidence. Just 1,513 ratified greenfields agreements were established between 2006 and 2010[1] (May, 2012), compared to 862,993 new organisations registered as trading entities in the six year period 2005-2010. Greenfields agreements may be only a small proportion of greenfield sites, but many "new organisations" may be either existing organisations rebranded or shells established for tax or related purposes. Quantitative investigations of genuine new organisationsin some senses a subcategory of the much richer literature on SMEsis an avenue for exploring the innovation in ER that arises from, or accompanies, a fresh start.
Such an exploration may also shine light on more passive explanations of union absence, as new organisations that are "old" firms masquerading as new (which may be the case with franchises (Kellner et al., 2014), replicator sites (Leopold and Hallier, 1997), or simply relocated old workplaces) have no inevitable, organic path to unionisation. Data on ER in emerging organisations tends to be limited and fragmented with international evidence suggesting that, at least as far as SMEs are concerned, ER tend to the informal (Edwards and Ram, 2009;Marlow et al., 2010;Wilkinson, 1999).. New firms may adopt a process of "mimic opportunism" and take on schemes which appear to work well in older organisations. Equally, ER initiatives may be part of an agenda shaped by a dependency relationship between large and small firms, or from pressures exerted by other organisations in a supply chain network. Issues in this respect often revolve around perceived legitimacy as smaller firms become pressurised to adapt to large firm expectations and demands (Barrett and Rainnie, 2001;Harney and Dundon, 2006;Wilkinson et al., 2007a).
From our perspective a pre-condition for innovation, is difference. If one type of firm is no different in its ER practices to another type, then it cannot be said to be more innovative in this respect. So we must first examine whether there are differences between new and old organisations in ER, and then, if there are differences, assess whether they signify greater innovation by one measure or the other. We referred earlier to the banality of concluding in circular fashion that new firms will be more innovative in their HR practices because any HR practices introduced will necessarily be "new to the unit of adoption", but that does not limit the ability of a study such as the one described here to surprise. This study includes, for example, an investigation of the source of ER ideas, a relatively unexplored field within knowledge management.

1400
PR 46,7 Method Participating firms "New" organisations have elsewhere been operationallyperhaps arbitrarilydefined as being less than five years old (Morehead et al., 1997) with the five year cut off also used in OECD definitions of young firms (Cieslik, 2007). The Australian Bureau of Statistics definition of a "small" business includes those with 20 employees or less, and researchers have observed that there appears to be a "genuine break" from small business practices (Holmes and Gibson, 2001, p. 12) when a firm reaches the 20 employee mark. Five years also provides managers with an opportunity to develop and test ER strategies. This study examines firms with a minimum of 20 employees that are either less than five years old (we call them "new") or more than five years old (which we refer to as "old").
Dun and Bradstreet's Who's Who in Business 2015 Australian database was sampled for both "new" and "old" organisations with at least 20 employees. That minimum size threshold was adopted both for data quality reasons (the Dun and Bradstreet database seemed less robust for organisations smaller than that) and conceptual reasons (we did not wish to excessively conflate the issues of very small size and newness) (Muurlink et al., 2011). Two groups of organisations were extracted. First, a population was drawn of companies five years old or younger, which yielded an N (after pre-screening by Dun and Bradstreet) of 634. Initial calls to the 634 firms established a total sample of just 182, with the balance either older than 5 years or with fewer than 20 employees. Of these 182 firms, completed interviews were obtained from 131 firms, representing a response rate of 72 per cent. Second, a random sample of 4,521 older firms was drawn from the total Who's Who in Business database to provide a comparison group. Attempts were made to contact all 4,521 firms. Screening eliminated 1,141 firms from the sample. Over 10 per cent of the initial sample (469), for example, proved to have less than 20 employees, 8 per cent of numbers provided (361) were not a telephone line (deadlines, no answers, or fax/modem lines), half a per cent (22) were not a business, and in 6.5 per cent (293) of cases, ten attempts were made to establish live contact without success. Of the 3,376 remaining firms, 57 revealed at interview that their company was less than five years old, and their responses were reassigned to the "new" firm category. A 38 per cent response rate (1,228) was obtained from the balance. The process thus yielded 1,228 "old" firm interviews, and 188 "new" firm interviews.

Respondents
The study sought the participation of "key decision makers". Respondents were spread between CEOs, general managers or managing directors (36 per cent); HR managers (36 per cent); other senior managers (e.g. administrative or operations managers) 25 per cent; and other staff (3 per cent). Of the respondents, 54 per cent were females. There were proportionately more male owners of businesses, more men in positions of general managers, managing directors, CEOs and finance managers while there were more women in the roles of HR manager and office manager/administrative support. Although the response rate was relatively high, there were some methodological issues raised by the fact that, while we deliberately sought to speak with only the most senior managers of the firm, in some cases the task of assisting with the survey appeared to have been delegated to less senior staff. However, we did include "don't know" options to account for this, and a reanalysis of the data on key items suggests that seniority did not account for a significantly different pattern in responses on most issues.

Questionnaire design
A draft survey was compiled using questions present in the small workplace version of the AWIRS and the Victorian Workplace Industrial Relations Survey (VWIRS) supplemented with questions developed specifically for the survey. Questions were designed to elicit the participants' industry by The Australian and New Zealand Standard Industry Classification (ANZSIC) 2006 (Revision 1.0), size, casualization and geographical configuration (Is this the only workplace out of which your business operates? Is this workplace the headquarters of the business?). Limited demographics about the individual managers responding to the survey were also collected, including job title, gender and education. Additionally, the study examined influence (What proportion of key decisions in this company are made by you?", "Who do you share decision-making responsibilities at your company with?), self-efficacy (When I make plans in relation to the business, I am almost certain that I can make them work), social support (How often do you talk about business problems with your family and friends outside the business?), work hours and stress. In addition to modified questions from AWIRS and VWIRS, the survey included a number of questions designed to establish the degree to which the firm had implemented specific procedures such as formal performance management and bonus systems (see Tables III and IV), the most recent such change (see Table VI), and influences thereon. Our initial interest, as mentioned, was in identifying whether, and to what extent, there were differences between new and old firms and then, if there were differences, assess whether they signify greater innovation by one type or the other. We also asked some union-related questions, not as an indicator of innovation, but to identify the type, if it existed of any enterprise agreement, and to measure unions as an impediment to (or source of ) innovation and as one of several aspects of ER with which respondents' satisfaction was measured. The interview length guidelines of the The Australian Market & Social Research Society (AMSRS, 2004) were adopted, with the survey piloted to ensure it conformed to the 20-minute limit.

Analysis
Parametric and non-parametric statistical analyses were performed to assess the association between old and new organisations on a number of key variables, including size of organisation, perceptions of the market context, management background and ideas, recent changes introduced to manage their people, the sources for the most recent change introduced to manage their people, general sources of ideas on managing staff, ER techniques used, features of the organisation, and unionism. Where content analysis was undertaken of open-ended questions in order to recode into analysable categories, the approaches of four coders were agreed upon and standardised. In some cases (e.g. size and casual employment) where variables were severely skewed, analyses were performed on inter-quartile range values. Chi square tests of significance were used with categorical data while independent samples t-tests were performed when the dependent variable was continuous (metric). Most of the analysis is bivariate (because we want to consider differences between new and old workplaces on a range of dependent variables) but we also undertake a regression (Table V ) to rule out confounding effects of other variables.

Findings
Industry Table I shows the distribution of respondents' employers by industry. Approximately 14.8 per cent of the organisations participating in the survey were from manufacturing, 12.3 per cent from professional, scientific and technical services, 9.7 per cent from health care and social assistance, 8.5 per cent (120) from education and training (8.5 per cent), 7 per cent construction, 6.4 per cent from wholesale trade, and 5.1 per cent from retail trade and accommodation and food services. The other ANZSIC divisions each accounted for less than 5 per cent of employment each. Two additional categories were created to deal with 1402 PR 46,7 multiple responses; one for those who replied that they were both in wholesale and retail trade; and a final category for other organisations with more than one activity. There were some significance differences between old and new organisations by industry. Firms in agriculture, forestry and fishing, electricity, gas, water and waste services, and construction industry were more likely to be "new" and manufacturing firms to be "old".

Employment
New organisations, as expected, were smaller than large organisations. While 52 per cent of full-time equivalent employees in old organisations were in organisations with 60 or more employees, only 37 per cent of those in new organisations were in similarly-sized firms. While only 24 per cent of full-time equivalent employees in old firms were in organisations with under 30 employees, this was the case for 38 per cent of those in new organisations. New organisations were also more likely to employ casual staff than old organisations (82 per cent compared to 74 per cent) even though, if casuals were randomly distributed, the probabilities would be reversed.

Market context
Respondents were asked to gauge four aspects of their market: the degree of competition; their reliance on a small number of customers; market expansion, and market unpredictability. They were also asked to give an indication of the financial position of the firm in the previous twelve months. Three of the variables produced significant differences between "new" and "older" firms: new firms were more likely than old firms to be experiencing market expansion, and to be experiencing relative market unpredictability. The competition context did not appear to differ much between new and old firms. 100.0 Notes: **Significant difference between "old" and "new" at p o0.001; non-significant differences between "old" and "new" are not shown in data columns 3 or 4 Source: ANZSIC (2006) With markets expanding more rapidly, new firms were also outperforming old firms on the single indicator of financial performance (see Table II), which probably also reflected some survivor bias effects (old firms being more able to withstand several "poor" years than new firms).

Unionism
The most notable ER difference between new and old organisations was in unionisation. Union presence in the surveyed firms was addressed through two questions, probing the presence of union members, and the presence of union delegates. Older firms (42.2 per cent) were significantly more likely to have union members present than newer firms (30.7 per cent) ( po0.000), but the difference between older firms (27.2 per cent) and new firms (17.5 per cent) on union delegates failed to reach significance ( p ¼ 0.15).

Alternative voice mechanisms
A number of questions probed the range of non-union employee voice features used in the surveyed firms. Respondents were asked about the presence of a range of common (top-down) voice mechanisms (see Table III). New organisations were less likely to use formal staff surveys (a plausible resource effect), but on other mechanisms, no statistically significant differences emerged.

ER characteristics
To create a picture of the overall ER architecture of new and older firms, respondents were also given a dichotomous choice (present/absent) for a range of features, either of a direct ER nature or with close relevance to the field. There is striking symmetry in the results. This "no difference" pattern emerges despite statistically significant differences between new and older firms on some of the core corporate demographic variablesperhaps the most surprising aspect of this study. Stark evidence of the symmetry of old and new is presented in Table IV, which lists 22 features with direct or indirect ER relevance.

Percentage of respondents New Old
We face a lot of competition in the market in which we operate 45 49.5 Most of our sales are to a fairly small number of customer 20.6 17.8 The overall market in which we operate has been expanding strongly in the last two years 37.6 24.2** The market in which we operate is highly unpredictable from one year to the next 29.1 22.4* In the last 12 months, the business has been travelling well 45 36* Notes: *po 0.05; **p o0.001 Only one of these 22, the presence or absence of a formal quality control system, varies with the age of the firm to a significant degree. Only one other, the incidence of formal job descriptions for all staff, even approaches significance. While there might have been other, possibly innovative practices that were not captured by the survey the consistency of findings is remarkable.
One possibility was that the extent of difference between new and old firms is masked by confounding variables, such as organisational size, unionisation or market circumstances. So we ran some ordinary least squares regressions to test the impact of such variables, treating as our dependent variable the count of the number of ER features (listed in Table IV) that an organisation had. Table V shows the results. Crucially, it indicates that the non-significance of the difference between new and old organisations remained, and the size and direction of the coefficient was little changed, even after controls for size, unionisation and market circumstances were in place. The regressions provided additional illumination on some other relationships. The number of features was higher in large organisations, and in those with union delegates, as well as in those exhibiting three market characteristics: operating in markets that had grown strongly in the last two years, a diversity of customers, and driven to release innovative products. Despite these three market characteristics, the extent of competition or unpredictability in the market had no separate impact. We do not go into these other factors in any greater depth here because they are not the main focus of this article: our point is simply that controlling for other variables made no difference to this indicator of the effect of age of business on the number of ER features an organisation had.

Recent changes
In contrast to the mere presence of business features with an ER flavour, when respondents were asked about the "most recent change", a number of contrasts emerged between "new" and "old" businesses. New businesses were generally significantly more likely to have most recently introduced one or more of the list of relatively standard ER features than older businessespresumably since older businesses are likely to have already implemented these measures (see Table VI). Significant differences emerged in responses on a number of recent changes: the introduction of a business plan or other major restructure of organisation, vision, values or culture approached significance in this manner, while off-site training for non-managerial staff, the introduction of written HR policies, voice mechanisms, recruitment changes, and the introduction of informal staff events reached significance. In relation to recent changes, we also offered respondents an open-ended response if their most recent change did not fit comfortably into the given categories. We then content analysed these open-ended responses and re-categorised them into the range of features shown in Table V.  This last optionto allow respondents to reveal that none of the standard features constituted their most recent changeallowed us to analyse non-standard (and presumably relatively innovative) ER changes. In all, 606 respondents offered these "other" recent changes (for example the development of wellbeing or leadership programs), with the results indicating that, to a marginally significant extent, older organisations (44 per cent) might have introduced these "other" changes more than newer entities (37 per cent) ( χ 2 ¼ 2.955 df ¼ 1 , p ¼ 0.097). Again this would make sense in that old organisations have probably introduced the more standard interventions earlier in their lifecycle.
New firms therefore, were more likely to have recently introduced certain formal practices while older firms may have been more likely to offer non-standard ER changes as their "most recent" change. One possible interpretation is that it is older, not newer firms that are more novel in their innovation. However, there may be a structural explanation to this difference. New firms may not have had time to introduce all the "main menu" items essential to operating a firm's ER effectively, and thus may not yet have had the opportunity to turn to further refinements in ER practice, these "other" changes. If new firms are in the throes of adopting much of the standard ER features of older firms, then older firms, already equipped with the standard "bundles" of features, could be introducing other changes interpreted as being innovative. It is also possible that new firms are introducing "main menu" measures but in a manner modified to their stage in the business lifecycle, or in a leaner form reflective of their limited resources. In the depth and detail of implementation, which cannot be examined in a time-limited telephone survey such as the current study, there lies opportunity for true innovation. So practices could be qualitatively different from to the practices observed in older and larger firms.
Even though the "most recent" change was probed for, multiple responses were allowed, and it is interesting to examine briefly both the number who took the opportunity to give multiple responses, and those who gave no responsesthat is, claimed not to have introduced any "recent changes" in ER practices. A substantial number of respondents, 335 (23.7 per cent) were unable to give an example of a recent change. A statistically significant larger number of newer firms (16 per cent) than older firms (4 per cent) gave multiple responses to this question. This reinforces the impression that new organisations may be introducing changes more rapidly, not for innovation, but for catch-up.

Source of ER ideas
Managers were asked to consider from where they got their ideas for their most recent ER change. Again, there was remarkable congruence between the results for "new" and "older" organisations (see Table VII). For "new" organisations, the top three sources of "change ideas" were current colleagues, themselves, and previous colleagues, in that order, while for "older" organisations, the top three were current colleagues, themselves, and other businesses in the industry. Significant differences were found in only three categories of sources: new firms were more likely to be influenced by previous colleagues or "yourself", while older firms were more likely to point to "legislation and regulation". The greater reliance on previous colleagues may well point to respondents having arrived relatively recently from positions at other firms.
Managerial satisfaction with ER characteristics of the firm Finally, the survey asked a number of questions about managers' degree of satisfaction about various ER aspects of the organisation (see Table VIII), with higher scores reflecting greater satisfaction, and a score of 3.5 indicating a midpoint. The results show a surprising level of commonality between "new" and "older" firm responses on most questions. The one exception was that managers in new organisations were less satisfied with their own knowledge of ER, presumably because they have had less time to become aware through experience of ER in their organisation but have also been introducing changes there faster.

Discussion and conclusion
We posed the question, "do new organisations show evidence of innovative, that is different, practices when it comes to how firms handle their staff?" That is, are they at the "leadingedge" or alternatively, the trailing edge of ER practices? This study, while a large, representative national study with good response rates, is cross-sectional, with the inescapable disadvantages of a "single-shot" study. Though it cannot conclusively answer whether new firms rapidly converge on the same suite of practices that characterise older firms, the results offer some clear inferences. It appears that, within a short time, new firms come to largely resemble older firms, at least in terms of adoption of ER practices, though we have no data on the quality or depth of implementation. Setting aside industry composition, the major difference is that new firms are less likely to be unionised than old firms. There are several likely reasons: new organisations are smaller and more likely to make use of casual employees, both of which are likely to make unionisation harder, and unions have to identify and "break into" new organisations. But it does not appear that new organisations are less unionised because of the application of sophisticated, innovative management practices that make unionisation difficult, as there is little evidence that such practices are found more commonly in new organisations.
The reputation of "start-ups" as hives of innovation is not matched by realityat least not in ER. There are a range of factors that promote standardisation in ER, such as the impacts of regulation, product market conditions, unions, employer associations, law firms and management consultants, and this study takes place in Australia, which offers a highly-regulated industrial context. Moreover, it is rational for new firms to attempt to minimise costs by emulating practices proven in established firms. Institutional isomorphism, as Selznick (1996, p. 273) puts it, occurs because "organisations tend to model themselves after similar organisations in their field that they perceive to be more legitimate or successful". It might be the case that new, high-level technology may require new work practices, either as a function of the technology or driven by the need to attract or retain qualified staff, but there is no reason from our data to believe that new firms will be more focussed on introducing new, high-level technology than old firms. (If anything, the data in Table VI suggest a slower uptake in new firms, but the difference was non-significant.) Change in ER practices appears more rapid in new firms mainly because they are in the throes of adopting much of the standard ER features of older firms. Age and size are naturally correlated. Firms that reach a quorum of 20 staff in five years are almost by definition gazelle firms, and while the literature on the recruitment pressure placed on high-growth firms has escaped scholarly interest (an exception being Coad et al., 2011), it is inevitable that new, high growth firms face specific HR challenges in settling and training new hires. The "crisis of growth" may cause what Leonard-Barton refers to as "core rigidities" (Leonard- Barton, 1992;Muurlink et al., 2012). Nevertheless our findings show that the symmetry between new and old firms holds when size is taken into account. It is possible that differences between "new" and "old" firms may have been exposed if we had included no minimum size bar in our new firm sample.
There were no significant differences in the presence of flexible leave arrangements for staff, individual contracts for non-managerial staff, the use of performance reviews in making promotion systems, enterprise agreements, or performance-related bonus systems. On the evidence offered by this study, it seems likely that as new organisations age and grow, they will also begin to adopt, not always willingly, other characteristics of older organisations, such as union presence. In fact, the picture of symmetry between new and older organisations is such that it would appear new entities are not seeking to actively remake work practices, but are trying to replicate the practices of older firms. This convergence has been examined in the technology sector by Martin-Rios (2014) who found that active sharing of human resource management knowledge occurs in inter-firm knowledge networks; our study confirms that key sources of ER knowledge include industry bodies, suppliers, rivals and former business colleaguesahead of formal business education which accounted for only between 3 and 5 per cent of ER ideas.
Mainstream media (and consultants) have a tendency to focus on outliers as a sign of innovation. Two high profile examples in current times are Google and Uber. Uber faces a strong online and media backlash regarding HR practices within their head office (Dorney, 2017). Google on the other hand has been in existence now for more than 20 years and was first listed on Fortune magazine's best companies to work for in 2007some 11 years after start up and beyond our five year limit for new firms. What is perhaps more important, is that both firms have very innovative products yet the ER practices are the product of managerial decisions that are similar to countless other workplaces, but a series of chance events happened to go in favour of Google, Uber and the like (Kahneman, 2011). These organisations have innovative products, embracing new technology, at the right time and, to be frank, we love to read case studies of extreme success or extreme failure. It is questionable whether these organisations were truly innovative in human resource management practices in their first five years.
There is a tendency in both management and consultant circles to accept the truism that new firms are hives of innovation, and there is evidence when it comes to technological innovation that this assumption holds in reality (e.g. Brem and Voigt, 2009). In the field of ER, our study shows that when it comes to innovation in ER, at least, there may be differences in the stage at which ER practices are introduced (inevitable considering the different stage of firms are undergoing) and there may be a difference in the degree to which these practices are implemented (something this study does not address), but the two sets of firms have chosen essentially the same set of ER options from the menu. New organisations may create the impression of novelty through their speed of adoption of the same old practices.