Do small firms fare better without a professional human resource manager?

We examine the direct and indirect effects of human resource (HR) informality — that is, not having a professional HR manager in post — on employee outcomes in growth-oriented small firms. Drawing from literature of principal – agent relationship and trust, we theorise a moderated-mediating model between HR informality and unfavourable employee outcomes through owner-manager ’ s distrust during firm growth. Based on matched employer – employee data from 543 small firms and their 4853 employees in the United Kingdom, the empirical results show a positive and significant association between HR informality and owner-managers ’ distrust of staff, especially in multiple-site (often a manifestation of growth-ori-ented) small firms. Through this path, HR informality has a significant and adverse indirect effect on firm recorded staff voluntary turnover, absenteeism, number of dismissals and staff perceived collective trust in management, which counteracts its weak positive direct effect, thus leads to an overall negative impact. The theoretical and practical implications are discussed.


INTRODUCTION
Small firms are a major contributor to economic recovery and employment creation, but there has been a paucity of studies on their employment practices, especially, when small firms scale up (Allen et al., 2013;Nolan & Garavan, 2016).When growth takes place, small firms tend to lack a formal human resource (HR) system, as a result, informal employment practices prevail (Atkinson & Sandiford, 2016;Bauweraerts et al., 2020;Marlow et al., 2010;Rousseau et al., 2006).However, the impact of HR informality on employee outcomes during firm growth remains unclear.We aim to fill this void by examining the role of HR informality, typically manifested by not having a professional HR manager in a specified post (Atkinson & Sandiford, 2016;Singh & Vohra, 2009), on employee outcomes (staff turnover, absenteeism, dismissal and employees' collective trust in management), especially collective trust in management is the essence of cooperation needed when firms grow (Colquitt et al., 2013;Whitener et al., 1998).
Extant studies have reported both positive and negative impacts of HR informality on employee outcomes.Idiosyncratic employee relations in small firms are typically underpinned by so-called 'I-deals'.A personalised employment agreement negotiated between an individual employee and the employer predicated on the individual's pro tem value to the employer, and the negotiation skills and personal relationships of the parties.This individualised employment arrangement is of a nonstandard nature and tends to vary between co-workers (Atkinson & Sandiford, 2016;Rousseau et al., 2006).The intimate personal bond between an individual worker and the owner-manager has led to the heroic nostrum that 'Small is Beautiful' (Saridakis et al., 2013;Storey et al., 2010), but I-deals remain a source of inconsistency and potential conflict (Atkinson & Sandiford, 2016;Marlow, 2003;Marlow et al., 2010).The latter effect may prevent the development of trust relations and cooperation at the workplace, thus hampering firm growth (Allen et al., 2013;Arregle et al., 2007).We endeavour to capture this competitive effect of HR informality on employee outcomes through its counteracting direct and indirect effect (Xiao et al., 2018;Zhao et al., 2010) when firms grew in size and across sites.Our matched employer-employee data supports the proposed competitive mediating effect and makes three contributions.First, we add to the study of employee relations in small firms that start to grow.Second, we elucidate the mechanism by which HR informality can adversely affect small firms' growth.Third, we proffer a methodological contribution by using a matched and aggregated dataset to measure collective trust and to explore the competitive effect of HR informality on employee outcomes through econometric modelling.
We present this paper as follows: the theoretical framework rooted in both principal-agent and trust theories establishes why and how HR informality can affect employee outcomes in small firms.This is followed by research methods and the findings which lead to our discussion and conclusions.

Professional management and informality in small firms
When small firms grow, a high degree of internal cooperation is vital to ensure the effective operation of business functions, buttressed by a high level of trust between management and staff (Leana & Van Buren, 1999;Marlow et al., 2010;Wang et al., 2018;Whitener et al., 1998).Employing professional HR managers is a sine quo non in establishing mutual trust through the ostensible impartial and independent implementation of formal HR policies and procedures (Collins & Smith, 2006;Colquitt et al., 2013;Dirks & Ferrin, 2002).However, small firms tend to live in the shadows of informal employment management, typified by the absence of a HR professional in a specified post to implement formal HR policies and procedures (Bryson & White, 2019;Mayson & Barrett, 2006;Ram et al., 2001).Consequently, the so-called 'I-deals' employment relationship has been widely observed in small firms.That is to say, employment terms and conditions tend to be individually negotiated of a non-standard basis and 'differ, to some extent, from co-workers' (Rousseau et al., 2006, 978).This has continued in small firms despite the need to comply with a raft of new employment regulations in the UK (Atkinson & Sandiford, 2016;Marlow, 2003;Marlow et al., 2010).These I-deals are concerning since they further strengthen idiosyncratic social relations in small firms (Allen et al., 2013;Marlow et al., 2010).
Unsurprisingly, employment practices are found to be a mixed bag ranging from the family-like metaphor in some studies (Saridakis et al., 2013;Storey et al., 2010) to the other end of the spectrum with 'Bleak House' operations (Wilkinson, 1999) and arbitrary favouritism (Marlow, 2003).We acknowledge the potential positive direct effect that HR informality may have on individual employees, and further examine its indirect effect on employee outcomes of the whole workforce when firms grow.
HR informality, at its worst, leads to greater unregulated exploitative employment practices.The UK has a history of efforts to address this issue by regulating the employment relation through the advent of more formal procedures at work.Such efforts are epitomised by the creation of the Advisory, Conciliation, and Arbitration Service (ACAS) in 1975 (ACAS, 1980) and Chartered Institute of Personnel and Development (CIPD) (founded in 1913 as the Welfare Workers Association, and became CIPD in 2000).The latter has focused on delivering a body of 'professionally' trained and well-educated HR managers (Seifert, 2015) who are hired (mainly by large firms) to manage the formal processes and practices relating to employment issues.
The fundamental argument behind formalised procedures and professional cadres of HR managers in firms was that, ceteris paribus, arbitrary management in employment relations could create inconsistency, thus low trust, and litigations (Edwards, 2003).Therefore, more formal procedures operated by professionals would reduce the risk of arbitrary conduct, improve employment outcomes and trust relations, and thereby lead to both better employee outcomes (Berkery et al., 2017) and better overall performance (Heery, 2016;Sheehan, 2014).From this viewpoint, all firms, whatever their size, are urged to emulate the formality of their larger counterparts as role models, and much of this received wisdom was subsequently put into employment legislation (Ewing, 2008).
Despite the fact that 'formal' HR policies and practices have been widely taken up in small UK businesses (Forth et al., 2006), its impact on employment outcomes was found to vary from negligible (Storey et al., 2010) to significantly positive (Bryson & White, 2019).The effect of having a professional HR manager, often a sign of HR formality, on employment outcomes is ambivalent (Lai et al., 2017;Storey et al., 2010).On this note, the debate on whether or not to hire a professional HR manager has been well rehearsed, pointing out certain factors that militate against hiring a professional HR manager in small firms.Notably, resources tend to be limited (Kakabadse et al., 2020;Ram et al., 2001) and the management of HR tends to be considered less technical than other management functions such as finance or technology, which makes the appointment of a HR professional seem to be the least urgent (Marlow et al., 2010;Wynarczyk et al., 2016).Moreover, owner-managers prefer to stay 'hands-on' with labour management for as long as possible.This occurs even when the owners know that hiring professional managers across the board can improve business performance (Sundaramurthy, 2008).

Principal-agent relations and trust in small firms
Firms are started with founders and often employ family members at the early stage, who work together with a shared goal of increasing family wealth.The economic pathology of firm growth means that more non-family members join in for their expertise or skills.This changes the dynamics of association and ultimately trust in the management.The level of association between each individual worker and the organisation varies from family member employees to non-family member employees and with that, the degree of goal agreement diverges (Campopiano et al., 2020).The principal-agent problem in employment relations in small firms (Eisenhardt, 1989) means that the principal (owner-manager) cannot directly ensure that employees (the agents) are always acting in the best interests (goals) of the firm.For most non-family employees, working in a small firm is just a job to satisfy their own financial and social needs (Allen et al., 2013;Chrisman et al., 2012).Such 'outsider' staff may indulge in activities that go counter to the firm's efforts to improve performance, such as opportunism, incompetence, and underworking (Taylor, 1911).
Distrust, defined as the confident negative expectation regarding others' conduct (Lewicki et al., 1998;Kramer, 1999;Van de Walle & Six, 2014), may take root due to the conflict of interest between the principal and agents, and asymmetric information.Conceivably, the level of natural suspicion of the principal tends to be higher than that of a HR professional for three reasons.First, there is often unmatched level of commitment between owner-managers and staff, thus tension between employers and staff.Owner-managers, in many instances, are with high degree of identification with the firm, which can lead to unrealistic expectations of commitment from employees, if thwarted, reinforces the owner-managers' suspicions of the workforce.Second, the low degree of competence exhibited when dealing with complicated employment relations within their overstretched roles can negatively affect employment relations.With firm growth, management expertise in different areas is required, such as the need to handle increased employment regulations (working time regulations, equality issues, and national insurance compliance), which is often far beyond the 'comfort zone' of the owner-manager (Kakabadse et al., 2020).While for staff, it is just a job and they may know more about labour regulations to protect their rights than the owner-manager does and then make full use of their knowledge, as is often the case with Working Time regulations (Marlow et al., 2010).
Lack of competence in this regard strengthens the suspicions that managers-that is, owner-managers/non-HR professionals-have concerning their own staff being shirking or taking advantage of management (Lewicki et al., 1998;Marlow et al., 2010).
Third, the physical distance between different business sites-which is often an effect of firm growth-can magnify the inconsistencies inherent in informal HR practices.When firms expand into multiple-site operations, any inconsistencies in handling employment issues due to habitually informal dealings can result in known 'unfairness'.The 'I-deals' employment arrangement (Atkinson & Sandiford, 2016;Marlow, 2003) can travel faster between staff and sites due to the relatively dense workgroup networks in each location and employees' shared interests as 'agents'.Employees' responses or shared knowledgeable conduct may directly challenge the inconsistency of 'I-Deals' (Rousseau et al., 2006) and competency of managers (Lobel, 2018;Marlow et al., 2010).This can make an owner-manager feel "taken advantage of by staff", thus amplifying the distrust of staff.This reasoning leads to our Hypothesis 1a and Hypothesis 1b.Hypothesis 1a.HR informality is positively associated with owner-managers' distrust of the workforce.
Hypothesis 1b.Owner-managers' distrust of the workforce is higher in small firms with multi-site operations than single-site ones, that is to say, a moderating effect of firm growth by having multiple-versus single site.

The pathway between HR informality and employee outcomes in small firms
The conventional wisdom is that workplaces with formal HR policies and practices can limit or reduce idiosyncratic and arbitrary attitudes to hiring and firing, grievance handling, performance management and reward systems.On this point, it is noteworthy that the impact of formal HRM policies and procedures is subject to a 'fair' implementation.This is especially hard to ensure when I-deals prevail (Atkinson & Sandiford, 2016;Marlow, 2003;Rousseau et al., 2006).A recent study reveals that when formal HRM practices are initially introduced in Small and Medium Sized enterprises (SMEs), it often negatively affects employees' workrelated attitudes, for reasons such as inconsistency, alienation, and distrust.Only when SME owners invest in HRM practices (including using professional service or staffing), can it lead to the development of positive employees' attitudes (Bryson & White, 2019).In addition, a formal policy, for example on equality and diversity, that is not followed through can cause more discontent among staff than the absence of such a policy (Nishii et al., 2018).The attitude of owner-managers towards staff shapes the ways in which processes are set up and how HR policies are implemented and thus affects the intended HR practices in small firms.
We argue that, as expected, the presence of formal HRM policies and procedures is a necessary but not sufficient condition for improving employment relations (and trust between managers and employees) in small firms.The premise is that those managers who are involved in operating the procedures (both in the decision to trigger the procedure and in the ways in which the process is handled once formalised) need to be fair-minded towards employees.It is important to accept that the procedures are there for the deliverance of 'just' outcomes within the internal procedures applied, without fear or favour, to all staff (Feuille & Chachere, 1995;MacCoun, 2005).This is rarely the case in firms with informal employment arrangements.
The mind-set of owner-managers or non-HR professionals, who are in charge of employment issues in small firms, is inclined to manage people based on I-deals (Atkinson & Sandiford, 2016;Marlow et al., 2010;Rousseau et al., 2006).This is partly due to limited knowledge, skills or competence in managing employment relations (Allen et al., 2013;Marlow, 2003;Wilkinson, 1999), and partly due to the pressures to keep going and focus on economic outcomes without the headspace to plan and to put the process into place.Owner-managers/non-HR professionals still exhibit legal incompetence when dismissing employees, reflected in the high proportion of unfair dismissal litigation among SMEs (Lobel, 2018).Moreover, the lack of professionalism in the handling of employees' concerns can further worsen employment relations, leading to undesirable outcomes, such as high sick leave, high staff turnover and failure to retain valuable employees.
On the other hand, the need to nurture employees' collective trust in management becomes even more pressing when firms grow bigger.This is because trust in management can ensure cooperative behaviour in the production of goods and services (Fink & Kessler, 2003;Kramer, 2010).Employees' trust in management is rooted in three aspects: abilities, benevolence, and integrity (Mayer et al., 1995).A shared perception of trust in management is influenced by how managers execute HR policies (Collins & Smith, 2006).Most research has featured trust as a dyadic (supervisor-subordinate) personal relationship (Colquitt et al., 2013;Dirks & Ferrin, 2002).This power-embedded personal trust may lead to favouritism of some but damaging trust relationship of the overall workforce and this becomes worse the more informal the treatment of grievances (Jiang & Probst, 2015;Marlow, 2003;Marlow et al., 2010).Arguably, perceptions of injustice can spread quickly among staff due to the existence of dense social networks in small firms, which can adversely affect their trust in management.This leads to our Hypothesis 2a and Hypothesis 2b.
Hypothesis 2a.HR informality in small firms is negatively associated with perceived collective trust in management through owner-managers' distrust of staff (a mediating effect).
Hypothesis 2b.HR informality in small firms is positively associated with undesirable employment outcomes (high dismissal, sick leave and voluntary resignation) through owner-managers' distrust of staff.

DATA SOURCE AND METHOD
The national representative Workplace Employment Relations Survey (WERS) dataset is used for the purpose of this study.WERS is a series of surveys on employment relations in the workplace in Britain collected in 1980, 1984, 1990, 1998, 2004 and 2011.The main objective of each WERS survey was to provide large-scale, statistically reliable evidence about a broad range of employment relations and practices across almost every sector of the UK economy.The principal investigators were the Department of Business, Innovation and Skills (BIS), Advisory, Conciliation and Arbitration Services (ACAS), Economic and Social Research Council (ESRC), UK Commission for Employment and Skills (UKCES) and the National Institute of Economic and Social Research (NIESR).It follows a systematic random sampling procedure and data collected through face-to-face structured interviews with the most senior manager responsible for employment relations and personnel issues.Permission is also sought from the manager to distribute a selfcompletion questionnaire to a maximum of 25 employees in the workplace.If a union or non-union employee representative is present, an interview with them is also sought.The dataset thus includes three sets of large-scale observations; the WERS2004, for example, includes valid responses from 2295 workplace managers, 1070 worker representatives and 22,000 employees.
Questions relating to managers' trust towards staff are only available in the WERS 2004 dataset, which is the second latest available survey, 1 making this dataset unique.We explicitly focused on small firms employing between 5 and 49 workers in total, including single and multiple-site operations.WERS provided such information to identify small firms (Lai et al., 2017;Saridakis et al., 2013).In order to capture mutual trust between managers and staff, we used matched responses from the managers' and employees' questionnaires.The latter is used to construct staff shared trust in management in which we used 5-25 valid responses per small firm.This amounts to 543 small firms with 4827 associated employees in total.Therefore, this dataset provides a sizable and robust basis for testing our hypotheses.

Independent variable
Not having a professional HR manager in post has been generally agreed as a proxy for HR informality (Lai et al., 2017;Marlow et al., 2010;Saridakis et al., 2013).We consider two key aspects to capture HR informality from the WERS managers' questionnaire: job title and job remit.The extent of HR informality is constructed as a scale from 1 to 4, the higher the value, the greater is HR informality.We denote that 1 = HR post staffed by a HR professional manager or a personnel officer thus job role is to manage employment relations (18%, 100 firms); 2 = no HR job title but employment relations is their primary responsibility along with other roles (9%, 43); 3 = a manager designated for non-HR job title (production or marketing) while employment relations management is a peripheral duty (58%, 319); 4 = proprietor or owner who are in charge of employment relations (15%, 81).We also included a wide range of formal HR policies and procedures as control variables to capture the level of HR formal practices (see details in Table A1).This measurement follows the bundle used by Storey et al. (2010) from the same dataset.

Mediating variable: managers' distrust in staff
Managers were asked to rate items which indicate their attitude towards employees (1 = strongly agree; 5 = strongly disagree).We used three items to gauge their trust in staff (Kramer, 2010): 'given the chance, employees at our workplace sometimes try to take unfair advantage of the management' (reversely coded), 'employees here are fully committed to the value of this organization', and 'employees are led to a long term employment in this organization'.We first checked the validity and reliability of these three items to measure distrust.This generated eigenvalue >1.52, factor load-ing>0.61 and Cronbach's alpha >0.65, and therefore, it is acceptable for use as one construct due to the sensitive nature of the question (Kaplan & Saccuzzo, 2009).

Dependent variable (employee outcomes)
We measured both employment outcomes and employees' attitudinal outcome extracted from two questionnaires.The former are firm records in three accounts: staff absenteeism, measured by percentage of working day loss due to staff absence in the last 12 months, ranging from 0% to 60% with mean as 0.44 (S.D. = 6.94); number of dismissals, 79% of the total firms had no dismissal; 14% with 1; 5% with 2; the highest dismissal is 52; this gives a mean of 0.41 (S.D. = 2.39); and number of voluntary resignations, 4% of the total firms had none, 21% with 1; 15% with 2; 10% with 3; the highest number is 41, giving a mean of 3.08 (S.D. = 4.33).
Employees' attitudinal outcomes are captured by their shared trust in management in the employees' survey.This is measured by the aggregated degree of trust shared with sufficient consensus among members in an organisation (Collins & Smith, 2006;Fulmer & Gelfand, 2012;Kramer, 2010;Mayer et al., 1995;Wang et al., 2018).There are three items measuring individual trust in management in WERS 2004 (1 = strongly disagree to 5 = strongly agree): to what extent can management be relied upon to keep their promises; deal with employees honestly; and treat employees fairly.We first checked the validity and reliability of these items to measure trust.It generated good indicators (eigenvalue >2.43, factor loading >0.87, Cronbach's alpha >0.88).We then aggregated individual trust in management of 5-25 valid responses out of 5-49 employees per firm.Aggregation statistics were computed to validate the aggregation of individual trust perception in management at the organisation level.The Average Deviation (AD) index obtained was less than 0.48, which was well below the threshold of 0.87 (Burke & Dunlap, 2002).The r wgJ > 0.72 was above the conventionally accepted value 0.7 (James et al., 1993).We also considered the intraclass coefficients ICC (1) >0.13 and ICC (2) >0.72.Together these indices provided sufficient justification for aggregation of individual perceptions of trust to represent shared trust as an organisational property and differentiate small firms from each other on this basis.

Moderating variable
Firm site is measured as 1 = this establishment is one of a number of different workplaces (62%); 0 = single independent establishment not belonging to any other place (38%).

Control variables
These include firm age (years in existence), size (number of employees), sector and percentage of skilled workforce for its reported significant influence on people management practices in small firms (Lai et al., 2017).We also included the qualifications of managers who are in charge of employment issues as a control variable for the purpose of this study.Tables 1 and A1 provide more details.

Empirical analytical methods selection
Since HR informality can have both positive and negative effects on employment outcomes (White & Bryson, 2016), we intend to compare the direct positive effect it may have on employment relations with a negative indirect effect through managers' distrust of staff.This is captured by the competitive mediating methods we employed (Zhao et al., 2010).That is to say there are opposite signs of the direct and indirect paths but the overall effect can be zero, which tends to be negligible in Baron & Kenny's (1986) full mediation analysis (Hayes, 2013).We first test the significance of each path in order to rigorously capture the indirect effect (Preacher et al., 2007;Yzerbyt et al., 2018;Zhao et al., 2010) and then calculate the indirect effect and overall effect.
Our Hypotheses 1a and 2a depict a mediating model and the Structural Equation Modelling (Stata 13 SEM command) is therefore employed.Since the dependent variables in Hypothesis 2b (employee behaviour: number of turnovers and sick leave and dismissal) are non-continuous variables, to calculate each path, the non-negative binomial regression is used where data with a large proportion of non-occurrence, is zero.

Empirical findings
Table 2 presents correlations among the main variables.In the first column, it shows that most multiple-site small firms have no HR professional in post, and the direct effect between HR informality and our interested dependent variables are, as expected, weak.However, in the second column, the significant correlation between managers' distrust of staff and dependent variables are evident.In addition, managers' distrust of staff positively and significantly correlates with multiple-site operations, indicating an indirect effect of HR informality when firms expanded.A Variance Inflation Factor (VIF) test shows that the value is less than 2 (not exceeding the threshold value VIF > 4), so multicollinearity is not a concern for the regressions.

Hypotheses testing and results
Step 1 in Table 3 shows that firm growth (size and expansion in locations) is negatively and significantly correlated with collective trust in management.Hypothesis 1a predicts a positive association between HR informality and manager's distrust of staff.In Step 2, it shows that the proposed relationship is significant and positive (β = 0.29, p < 0.01).Furthermore, firm growth is significantly and positively correlated with managers' distrust of staff.Compared with the manufacturing sector, managers' distrust of staff is significant and negative in the financial, health, and education sectors.This could be attributed to the higher levels of skilled workers in these sectors.Therefore, Hypothesis 1a is supported.
Hypothesis 2a predicts that HR informality can negatively affect collective trust in management through managers' distrust.This is examined in Step 3 of Table 3.It shows a negative and significant correlation between managers' distrust of staff and collective trust in management (β = À0.17,p < 0.01).The indirect effect through managers' distrust is negative and significant (β = À0.05,p < 0.01) in the middle of Table 3.Since the direct effect is positive and negligible (β = 0.01, p = n.a.), this results in an overall effect of being negative and significant (β = À0.04,p < 0.01).The bootstrap with 5000 replicates supports the same results.This indicates that HR informality contributes to an overall negative and significant effect on collective trust in management.
Hypothesis 1b predicts an increased manager's distrust of staff with HR informality after firms grew.In the last column in Table 3, we include an interaction item between HR informality and multiple-site operation; the coefficient is significant and positive (β = 0.31, p < 0.01), indicating that the level of managers' distrust is amplified when firms operate in multiple sites.In Figure 1, we further explored this moderating effect visually: compared with professional HR managers in post, owner-managers' distrust of staff is significantly conspicuous from singlesite to multiple-site operation.This provides empirical evidence to support our Hypothesis 1b.
Hypothesis 2b predicts that HR informality can lead to negative employee behavioural outcomes (resignation, absenteeism and grievances) through managers' distrust of staff.We investigated the relationship through three negative-binomial regressions in Table 4.The path analysis included both the direct and the overall effects of HR informality on employees' outcomes.Although HR informality has an insignificant overall effect (second column of each specification) on employees' resignation (β = À0.13,p = n.a.), dismissals (β = 0.13, p = n.a.), and sick leave (β = À0.04,p = n.a.).Managers' distrust, however, is significantly associated with a higher number of voluntary staff resignations (β = 0.13, p < 0.01), higher number of dismissals (β = 0.33, p < 0.01), and a higher number of staff sick leave absences (β = 0.24 p < 0.01).Together with step 2 in Table 3 which shows the significant correlation between HR informality and  managers' distrust, Table 4 shows the indirect effect of HR informality through managers' distrust has worsened the overall effect of HR informality on the above employee behaviour outcomes by 35%, 22% and 29% respectively (calculation based on model 5 by Hayes, 2013).Since each individual path of the indirect effect, take HR informality-> managers' distrust-> staff resignation for example, is statistically significant, this has ensured the robustness of the indirect effect calculated (Yzerbyt et al., 2018).It thus provides evidence to support Hypothesis 2b.We further carried out the Medeff 5000 test for a simulated sample with 5000 observations that confirms the findings.That is to say, the overall effect of HR informality on employee behaviour is insignificant, because the indirect effect of HR informality through managers' distrust has counteracted the beneficial direct effect that HR informality may have.In some cases, this is very striking, take the percentage of working days lost (absenteeism), for example, the direct of HR informality is negative (β = À0.04,p = n.a.); after taking into account managers' distrust of staff, its overall effect becomes positive (β = 0.06, p = n.a),leading to a higher level of staff sick leave.The same applies to the number of staff dismissals, the direct effect of HR informality is small (β = 0.13, p = n.a), but after including managers' distrust of staff into the equation, the overall effect has worsened (β = 0.21, p = n.a).The results support a competitive mediating effect (Xiao et al., 2018;Zhao et al., 2010).

DISCUSSION
The study intends to decipher the constraints on small firms scaling up from the employment relations' perspective.Theorised from the principal-agent and trust literature, we explored both direct and indirect impacts of HR informality on employee outcomes in growth-oriented small firms.This enables us to reveal an overriding detrimental effect of HR informality through amplified owner-managers' distrust of staff when firms expanded into different sites.In this regard, this study adds to the literature on the complexity of employment relations in small firms (Atkinson & Sandiford, 2016;Marlow, 2003;Wilkinson, 1999) under growth conditions.By examining the mechanism of HR informality on employee outcomes through comparing its direct and indirect effects, the empirical results (in Figure 2) support a significant and negative indirect effect of HR informality on employee outcomes through owner-managers' distrust of staff.It appears that the detrimental effect of distrust overshadows any positive effect it may have.The study thus sheds some light on the inconclusive findings of HR informality in small firms (Lai et al., 2017;Saridakis et al., 2013;Storey et al., 2010).
HR informality tends to be natural to ownermanagers in small firms as a way of existence (Rousseau et al., 2006) and was found to stay when complying with new regulatory requirements, such as flexible working arrangements, in the UK (Atkinson & Sandiford, 2016).Since small firms are not 'small big firms' (Welsh & White, 1981), the growth path is a fundamental example of a quantitative change leading to a qualitative change in the power relations at work exemplified by the shift from informality being an asset to it being a detriment.HR informality, such as the I-deals (Marlow et al., 2010;Rousseau et al., 2006) in different operational sites, are hard to stop unaided.The inconsistency generated from individualised employment arrangements will be challenged by both legal and trade considerations including employment law and regulatory guidelines.The appointment of an HR specialist signals to employees, customers, and outside bodies the formal intention to abide by the rules and regulations.It is the first step to uproot employment practices inconsistency through eliminating new I-deals, implementing employment regulations with professionalism, and gradually reducing self-contradictions caused by old personalised employment deals among staff.Thus, it tends to improve employee outcomes in growth firms.
A collective trust in management among employees, however partial and contingent, is a necessary condition to ensure cooperation when firms grow.Small firms that benefit from the close-knit proximity and communication through HR informality may attain a higher level of personal dyadic trust between an owner-manager and a specific employee (Storey et al., 2010), but staff collective trust in management can suffer, especially so when firms expand into different sites.As shown in Figure 1, distrust of staff among owner-managers, compared with HR professionals, increases significantly from single-site to multiple-site operations.The challenge to trust, embedded in HR informality, when firms scale up are compounded by employment complexities, which go beyond managers' skill sets and may intensify owner-managers' distrust of staff.This distrust can spread into dealing with employment issues, thus having a detrimental effect on cooperation and employee outcomes.
Finally, our methodological contribution includes using matched datasets and adopting competitive mediating modelling.This makes the empirical results of this study more robust.We used managers' attitudes towards employees to match up with a sizable set of employees' attitudes (5-25 staff per firm) towards managers to test the proposed hypotheses.This matched employeremployee data can better capture collective trust in management.The information on managers' distrust of staff is very sensitive, and as such, this effect on employees has not been explored before with large quantitative data.Studies in this domain have typically been undertaken in the laboratory (De Cremer et al., 2018) or by limited interview samples.Using a consensus among a number of employees in small firms to measure collective trust in management makes the outcomes more convincing than those studies that derived from an individual rating system (Pielsticker & Hiebl, 2019).The proposed effect is further strengthened by HR recorded employee outcomes: number of voluntary resignations, sick leave and dismissals in the previous 12 months.
Adopting the competitive mediating modelling in the study of employment relations allowed us to capture the complex social relations developed within HR informality.This is down to acknowledging both positive and negative effects of HR informality on employee outcomes in small firms.In so doing, this study casts light on prior inconclusive findings in this regard (Lai et al., 2017;Marlow, 2003;Saridakis et al., 2013;Storey et al., 2010).

Managerial implications
In the UK, small businesses (0-49) account for 99.3% of the total business population, among them, 76% of businesses did not employ anyone aside from the owners.There are 1.15 million small businesses with 1-9 employees, but only 211,845 businesses with 10-49 employees (Department for Business, Energy & Industrial Strategy, 2020).We focussed on the latter group plays a vital role in job creation and economic growth.Among these firms, there is greater need for professionalism to ensure a 'fair and open' workplace climate-a scenario where the impact of HR informality needs unpacking.The adverse effect of HR informality on employee outcomes (sick leave, staff voluntary turnover and dismals) which are particular concerning given the scarce resource (including manpower) in small firms and it is susceptible to legal cases of unfair dismissal (Lobel, 2018), thus the impact on its productivity (Berkery et al., 2017) and performance.
A specified HR role (full-or part-time) staffed by a HR professional in small firms can institutionally ensure implementation of appropriate formal employment policies, which helps develop consistent employment practices within the legal and professional framework when firms have an increased geographical spread.This appointment can be vital despite pressures on scarce resources as we showed in this study.In addition, training initiatives for senior managers or owner-managers of small firms should focus not only on developing technical skills and business competences (Flatten et al., 2011) but also on addressing professionalism related to consistent employment practices among employees and to minimise the I-deal practices (Atkinson & Sandiford, 2016;Rousseau et al., 2006).

Limitations and avenues for future research
The sensitive nature of mutual trust between managers and employees makes it difficult to capture, especially managers' attitudes towards their workforce.WERS2004 provides not only managers' attitudes towards staff, but also allows us to match them up with up to 25 employees per firm and HR recorded employee outcomes.Such matched information and a sizable proportion of employees in small firms ensures the robustness of our analysis.There are, however, certain limitations.First, it would be ideal to have responses from more than one manager to gauge managers' trust in staff, just as we consider staff's trust in management was based on a sizable number.This is not possible in the current dataset; however, we view our approach as acceptable since we studied small establishments (5-49 staff) where the total number of managers is very small (often just the ownermanager).Second, the data collection took place in 2004.Yet, this is the only dataset that has questions that measured trust towards each other between management and employees.Importantly, the passage of time does not affect the contiguity of influence in this study, nor qualifiers/positive effects such as cooperation derived from trust in management and firm growth.Future studies may find more innovative data collection methods to overcome the above limitations and to improve our understanding of employment relations in small firms when scaling up.
Since the small firm sector is heterogeneous, future studies are needed to examine a range of employment practices in growth-oriented firms operating in different industrial sectors, with different ownership structures, workforce composition, and different national culture.Moreover, new forms of working, such as working from home/anywhere, present another avenue to explore employment relations among such firms.

CONCLUSIONS
This study explored the effect of HR informality on employee outcomes in growth-oriented firms, which are one major sources of innovation, employment creation and economic growth in the United Kingdom.The empirical findings support an overriding negative effect on employee outcomes resulting from managers' distrust in the workforce due to the absence of a professional HR manager in post.The effect is worse when firms expand into different sites.Since the appointment of a professional HR manager in small firms tends to be after the recognition of failing employee relations, this study provides evidence that such a time lag in the abandonment of HR informality can lead to high staff sick leave, legal cases (such as unfair dismissal) and low shared trust in management.All of which can impede growth.
This study, therefore, makes three main contributions: first, it reveals that employee relations tend to become more complex when firms start to grow, while distrust rooted in Principle-Agent relations may outweigh the positive features of social relations ascribed to HR informality in small firms.Second, we set out the mechanism by which HR informality may turn to a constraint on firm growth.Thirdly, we submit a methodological contribution by applying the competitive mediation model to capture the counteracting effect of HR informality on employee outcomes.

ORCID
of staff is moderated by firm growth T A B L E 4 Regression results on staff voluntary turnover and dismissal (negative-binominal regression) *** (H1a) HR informality Staff collecƟve trust in management -0.04(-0.05***)F I G U R E 2 Conceptual model with empirical outcomes.Note: bolded paths depict the central indirect effect of interest (**p < 0.05, ***p < 0.01) DO SMALL FIRMS FARE BETTER WITHOUT A PROFESSIONAL HUMAN RESOURCE MANAGER?