“Performance of SMEs: Literature synthesis of contingency models”

The fundamental issue of performance of SMEs has been intensively analyzed by aca- demia mainly through performance models. Performance models are largely focused on predicting failure and have intrinsic limitations. The present literature synthesis proposes an alternative perspective from seemingly unconnected empirical studies, which share the performance of SMEs as a common criterion: contingency models. Such perspective emphasizes the positive facet of SMEs results and complements the current academic standard of prediction models. This study performed integration and conceptual classification of 99 articles cataloged as contingency models and published by reputed journals from 1999 to 2019. The evidence from the integrated empirical literature revealed an ample and diverse set of supported variables that explain a positive economic outcome in small ventures. The analysis and classification of the articles derived 7 sections, 24 categories and 131 supported variables, which are intended to help improve the scholarly and professional analysis of SME performance due to the inherent viability of the integrated articles, proving a benign economic output. This view contradicts the current paradigm of performance models, which is insufficient to predict or explain how to improve SMEs results. The integrated elements developed by this study can work as a framework for the academic research of SMEs performance, and practical guidance for SMEs’ managers. This synthesis sets an agenda for further academic research in the field of SME performance, specifically in the areas of meth- odology, operationalization of performance, and empirical research.


INTRODUCTION
The performance of small and medium enterprises (SMEs) is of primary importance for the economy and society in general. As such, it has been a matter of intense scholarly scrutiny (Halabí & Lussier, 2014). Most of those studies have delivered their analytical efforts over one overarching format: performance models (Maes et al., 2005). Predominantly of prediction of failure (Amankwah-Amoah, 2016) or survival (Cader & Leatherman, 2011), performance models are the standard in the academic analysis of SMEs' economic output. However, performance models have important limitations like covariate selection (Gupta et al., 2018), are concentrated on the "negative" side of performance (i.e., failure), and leave aside an ample number of elements that can influence the positive performance of SMEs.
Alternatively, there is a vast, yet dispersed number of investigations named "contingency models", which contain substantive elements of improved performance in SMEs. Contingency models (Wiklund & Shepherd, 2005) are a wide set of empirical studies, from a varied number of disciplines that, although seemingly unrelated, have a common denominator: SME performance. This study integrated such investigations, which represent empirical evidence of a rich number of

LITERATURE REVIEW
The purpose of this literature review is to provide a picture of the previous scholarly work addressing small firms' performance: performance models as the dominant, yet limited paradigm and contingency models as the alternative though disperse perspective. Organizational performance has been a matter of intense academic research. The literature defines it from the perspective of its outputs and indicators, as pointed out by Santos and Brito (2012, p. 98): "firm performance is a subset of organizational effectiveness that covers operational and financial outcomes". Porter (1981) on the other hand employs three basic elements to define performance: profitability, cost minimization, and innovation. For Richard et al. (2009), performance "encompasses three specific areas of firm outcomes: financial, product market and shareholder return". This study focuses on SME performance due to their global importance and their sizable relevance at an economic and social level (The World Bank, 2018).
The literature has approached the phenomenon of SME performance from one front: performance models. They have dominated the landscape of the scholarly efforts to analyze the economic output of small businesses (Gaskill et al., 1993;Gupta & Gregoriou, 2018;Maes et al., 2005). They have a predictive proclivity, normally focused on either failure, survival, or success. Performance models normally occupy discrete/continue time options to estimate the moment of bankruptcy, which is key in predicting failure. They employ covariates within their multivariate models as their funda-mental indicators to determine whether the firm can fail (Gupta et al., 2018). The majority of performance models are concentrated on one pole of SMEs performance: failure or survival of SMEs (Abdullah et al., 2016;Cader & Leatherman, 2011), due to the generalized concerns of reducing failure rates (Amankwah-Amoah, 2016), as opposed to success prediction models, which are less common (Lussier & Pfeifer, 2001). This is presumably due to the perceived priority of preventing insolvency (Halabí & Lussier, 2014).
The core purpose of SME performance models is to supply reasons that explain and help predict business cessation. The common reasons delivered by performance models are working capital management, poor planning, deficient internal processes, lack of key resources, and lack of capital in the internal or resource-based facet of prediction models ( Watson, 1998). The contribution of failure prediction models to the literature is relevant: they have helped understand the important although limited reasons for the negative performance of small companies, assisting managers in trying to prevent bankruptcy.
However, most prediction models posit two major limitations. Firstly, they are focused on the "negative" side of the performance spectrum: the models explain what is behind a failure. As valuable as this is to avoid bankruptcy, in a modern demanding context and the higher entrepreneurships' ambitions and expectations (OECD, 2019), that might not be enough. It is necessary to provide SMEs' managers with an integrative guidepost that can explain "positive" performance: a favorable economic/financial output. It is equally necessary to contribute to the extant literature with a more global and comprehensive theoretical perspective for analyzing SMEs' output.
Secondly, performance models provide a rather limited typology of elements that can work as a theoretical framework for appraising organizational performance as commented above. This condition has, in turn, two manifestations. One is the limitation of the covariate selection: the nomination of the independent variables within the common multivariate failure models. This variable selection has been reported as "atheoretical" (Gupta et al., 2018). There is seemingly no proper theoretical underpinning for the selection of the covariates used in the models to test their relationship with the probability of default. The second manifestation is that they are "context-bound" in their majority (Abdullah et al., 2016;Lussier, 2001). The empirical data employed by the empiricists to test the models were extracted from specific locations or at specific industries. That is of value for the local conditions of the potential recipients of the studies but is limited for wider audiences due to their restricted generalization. Performance models play again their theoretical role in explaining one side of SME performance but might be insufficient in providing a more complete explanation to understand how to leverage performance.
On an alternate side, contingency models embody empirical studies that have investigated the positive side of the phenomenon of SME performance: the empiricists in these studies have proposed several precursor variables or "drivers", which have proven a direct positive effect on the results of SMEs ( Contingency models confront a rather challenging task in the academic investigation of SMEs: the intrinsic difficulty of replicating a meliorated outcome. The empirical studies representing contingency models are dispersed across the literature and apparently unconnected. This is arguable because they belong to different disciplines. Nonetheless, they share an uncontroversial common denominator: SME performance. They represent a rich and diverse source of elements that have the potential to explain leveraged results in SMEs. This study proposes employing those elements derived from contingency models to compose an integrated theoretical guidepost for the analysis of positive output in SMEs.
This review examined the current theoretical view for the analysis of small firm's performance and disclosed how it is largely focused on predicting their failure, which poses important limitations in explaining how to leverage the output of SMEs. It also revealed the alternative perspective of a wider number of investigations that presumably explain positive or benign results of small venues. In light of this background, it is important to complement the current academic investigations on SME performance with theoretical insights that can go from survival or bankruptcy avoidance to guidance on how to improve their results and take small entities to a more favorable operational condition.

GENERALIZATION OF THE MAIN STATEMENTS
The method followed by this study started with the compilation and analysis of the extant literature that has studied SME performance. The operator's "performance", "financial performance", "SMEs", "small business", "medium enterprises" were applied onto the major academic databases (i.e., Google Scholar, Scopus, Emerald, Jstor, Springer link, ProQuest, Ebsco, Science Direct, SSRN, and ResearchGate), prioritizing reputed journals specialized in SMEs from 1999 to 2019. The search unveiled a robust piece of SME literature, which had employed performance as a common criterion. A systematic literature review (Becker & Jaakkola, 2020) was performed of the seemingly unconnected empirical literature that appeared to represent a global source of elements related to SMEs' positive outcomes.
The literature search derived a total of 125 articles meeting the criterion, 68 of which came from reputed journals. 16 conceptual papers were removed to focus only on empirical material. 10 additional articles were excluded because this study could not get access to the original articles. The final sample was composed of 99 articles (Table  A1, Appendix A). The results were mapped and organized under the aggregated criteria "external and internal", accounting for the origin of the independent variables from their level of analysis. This is in agreement with the most common theoretical frameworks occupied by the very same articles in the sample. Amankwah-Amoah (2016) in their points of indexation or "sections" for appraising SME performance. The analysis of such integration returned specific performance categories.
The integrated literature of contingency models is reported in Table 1. It shows the summary of the sections utilized for integrating the literature divided by the level of analysis and the categories resulting from its analysis. As mentioned above, the sections are the points of indexation invoked from extant performance models. The categories are the result of the analysis performed on the integrated literature. They were extracted from the original articles and represent the different disciplines or domains employed by the empiricists. In this study, those categories allowed classifying and describing different variables. For example, Leonidou et al. (2016) in the study on environmental determinants within the literature on SMEs' environmental issues reported that environmental regulations (defined as a supported variable in this study), which was part of the broader business environment (extracted here as a category), had a positive impact on SME performance.
Utilizing the categories derived by the analysis of the integrated literature, this study classified the varied and vast number of independent variables or "drivers". Table 2 and Table 3 introduce the drivers within their sections and categories. Those drivers are the supported precursors in the original articles: the independent variables proposed by the studies that were positively related to the dependent variable performance. The unsupported variables were not included, in consistency with one of the core objectives of this study of focusing on the positive side of SMEs' results.
The numbers represent supported independent variables within the studies. The total of variables may vary from the total of studies, as each study may have more than one supported variable.
The numbers represent supported independent variables within the studies. The total of variables may vary from the total of studies, as each study may have more than one supported variable.
The present study performed the above analytical procedure to attain the study's goals and in response to the number and diversity of variables resulting from the sample. This study took into consideration the level of analysis ("origin"), the connection with the current theoretical framework ("sections" from performance models), and the context (the constructs, domain, or discipline of the original studies ("categories")) to integrate and classify the scattered literature onto an actionable theoretical point of reference.
These concepts (the level of analysis or origin, the points of indexation or sections, the constructs or categories, and the variables or drivers) are built upon the solid portion of evidence from the empirical literature that is contained in contingency models. They constitute thus an alternative integrated framework for the analysis of SME performance, which can provide further explanations of a leveraged economic outcome. Such shifting theoretical perspective can also help SMEs' managers to improve their results through implementable factors and set a path for further research.

Internal processes
Organizational policies --Internal procedures

DISCUSSION
This synthesis attempts to provide an alternative theoretical perspective to assist in the scholarly analysis of SME performance with an emphasis on the positive or leveraged economic/financial outcome of small entities. It also aims at helping managers with feasible elements that be used to improve the results of their small businesses. A thorough ex-amination of the extant literature was performed to confirm the dominant position of failure prediction models and to unveil a robust piece of empirical literature that, although dispersed, had the potential to provide sufficient evidence to improve our understanding of how SMEs can perform better.
The origins, sections, and categories introduced in Table 1 represent a compendium of the classified empirical evidence that can explain a positive output in SMEs. As introduced earlier, sections are indexation points taken from extant performance models. Categories were extracted from the sample and allowed classifying and better describing the different types of variables. So then, the first internal management section was assigned the categories managerial skills, competences, managerial actions, and strategy. The management section represents all the categories related to the personal abilities employed by SMEs' owners/managers in the managerial operation of their business, outside financial matters. The management section accounts for 35 out of 94 total internal variables (  In the external origin, the section environment, which embodies the categories business environment, environmental actions, competitive conditions, and economic circumstances, accounts for 51% of the total drivers under the exogenous level of analysis. This category incorporates external factors related to the physical context in which the firm operates and can have an impact on its output, as well as actions taken to directly cope with it. Such variables as environmental responsibility or market dynamism were elaborated. This is in line with common theorists invoked by the empiricists, in regard to the external causes of improved performance (Hunt & Morgan, 1995). They are equally in consonance with their counterpart in performance models, where a hostile environment can explain failure (Amankwah-Amoah, 2016). An example of the category came from Qian and Xing (2018), who uncovered a positive association between environmental responsibility and the financial outcome of 138 small venues in Australia.
The external section access to financing composed by the categories easiness to getting funding and readiness of financial services represents the exogenous conditions that small firms face in their local context to acquire financial resources. Performance sections and categories can act as a point of reference in further studies on SME performance to classify or to propose distinct precursors of performance, because they represent the very constructs, domains, or disciplines of different reputed investigations from around the world. They carry thus methodological viability and theoretical pertinence from the underpinnings employed by the empiricists, who were published by reputed journals specialized in SMEs. In this study, sections and categories were distilled and developed to classify and describe the articles' supported variables: the drivers of performance.
The drivers of performance captured in Tables  2 and 3 impersonate the independent variables proposed by the empiricists, which were positively supported by their statistical analyses. They, therefore, entail proven explanatory variables of improved economic outcomes in SMEs. The number of variables (94 in the internal origin and 37 in the external one) differs from the total number of articles in the sample because one study can have more than one supported variable. Out of the total 131 supported variables, 46 are repeating -occurring more than once-and 85 are different. The fact that the internal origin exceeds the external level of analysis concords with the most common theorist in the field of firm performance, arguing that endogenous resources are more common predictors of performance than the exogenous ones (Barney, 1991). The methodological profile of contingency models shows a pronounced concentration on quantitative methods, suggesting that the level of previous investigations on the subjects was reasonably sufficient. The dependent variable is changelessly SME performance and is measured either by "dynamic" economic units or by "static" financial indicators. The multi-national character of the inquiries is patent: 41 different countries in 99 studies. This denotes the extent of the investigations on the performance of SMEs, which took place across countries and cultures and seemed to be supra-location.
The aggregated analysis of Tables 2 and 3  The number and diversity of the drivers noticeably contrast with those of performance models, which are more limited. In like manner, the updated variables in contingency models are barely seen in the failure prediction literature. Performance drivers can enlighten the scholarly analysis of SMEs and the managerial actions to attain improved results. They can work as updated variables for further study and as an actionable list of precursors to serve in the aggregated analysis of a specific industry or geographic region. They can also be used in combination to address the variety and complexity of subjects that need to be considered in the appraisal of performance.
As appealing as contingency models might appear to researchers interested in investigating the field of small ventures, their dispersion, and other conditions come as limitations that need to be considered. The fact that contingency models are not connected may call for a conceptual integrative approach but could also signify that they are indeed not classifiable under the same umbrella. All the studies presented here belong to dissimilar disciplines, and what they have in common (performance as the prevalent criterion) might only be the necessary and only point of reference to appraise the effect of the precursors. The empiricists were genuinely preoccupied about the effect of their independent variables, but within their own fields, not necessarily in the pursuance of performance in itself.
Another limitation of contingency models and therefore of this study is the lack of a standardized methodology. Although there are some clear patterns in the research approach like the control variables, more academic rigor is necessary to ensure that the appraisal of the outcome is sufficiently insightful. The claim of causality is also limited by the nature of the drivers: although the models may have an explanatory orientation, the elements for causality might not be present in all the precursors. Some of them have inherent temporal precedence, but some others may only have positive associations and alternative explanations. Finally, the concept of positive performance may be an idealistic pursuit, as many small businesses are in reality facing numerous constraints that might threaten the continuity of their endeavors.
Analyzing the performance of SMEs from the academic perspective is not an easy task; nor to manage a small firm. There are simply too many fronts: marketing, finance, sales, competition, regulation, customers, and quality. Moreover, the list goes on for businesses to take care of. It is necessary to employ a framework that is sufficiently ample to reasonably cover the different angles of modern business. Contingency models can support such a framework. This literature synthesis suggests that their integration and classification can supply business diversity and theoretical pertinence. It is proposed to employ the origins, sections, categories, and drivers developed in this study to assist academic studies. The categories serve as a classification point and the drivers as proposed constructs in empirical research.
The rationale of proposing contingency models as a framework has been discussed throughout this study: contingency models embody a map of empirically tested factors for tangible positive results, linked together by the common denominator of economic/financial outcome as the literature's criterion. Overall, the variables clearly reflected the main concerns of the scholars. This represents an updated outlook because the focus was put on the subjects that encircle the most common trends, theoretical views, and impactful issues in their disciplines and countries. In contrast with traditional performance models, this proposed integrative framework does not employ a prediction tool: it has rather been derived from the actual empirical findings in contingency models. Unlike failure prediction models, which are naturally context-bound because the authors tend to build their constructs using locally collected data as referred above, the present study consolidates several already-supported predictors from varied fields and different geographical contexts and industries. The supported drivers presented here, with their respective categories and sections, have the inherent theoretical and methodological underpinnings of known authors and journals. They represent a compendium of a global, diverse, and viable set of elements that can further explain positive SMEs' results.
From the point of departure of the scope of this study and the academic problem addressed, the need for further research is evident. The scope of this conceptual study was enunciative: to integrate and classify a more ample set of elements that can improve our understanding of leveraged performance in SMEs and assist in its analysis. The contextual problem of this investigation was the limited perspective implied by the dominant framework of performance models, as well as the dispersed condition of contingency models. This synthesis detected four clear areas for further conceptual and empirical research: the methodology employed by the empiricists in studies related to performance, the measurement of performance as a dependent variable, further integration of studies with performance as a criterion and intended empirical research on performance.
The methods employed by empiricists when analyzing the impact of distinct variables on performance are pivotal for the reliability of their results. However, given the dissimilarities across the studies, more conceptual work is necessary here to guide further empirical studies. Especially regarding the investigation of the methods employed, the analytic techniques, and the conditions of the mediators and moderators. In the same line, a more critical area is calling for further theoretical study: the measurement of performance as a criterion. The present study found sound contributions ( Further theoretical integration of empirical investigations, which employ performance as a dependent variable to assess the impact of various precursors, is required in the literature of small companies. It is suggested in this synthesis that, regardless of the discipline of the studies, their unequivocal criterion of performance might constitute a uniform set of elements for better un-derstanding the positive outcome of SMEs. In that line, some authors have agreed on the "ultimate" role of performance as the dependent variable par excellence on SMEs research (Pucci et al., 2017). But does its presence suffice to consider all those studies as a valid set of homogeneous elements to independently analyze the economic outcome of small venues? Additional conceptual and integrative work can bring more clarity to determine whether or not is feasible to hold such an argument.
Additional empirical research is required as well.
Research that is explicitly intended to analyze the performance of small venues itself, beyond models and any pre-determined "type" of outcome that could constrain the precursors. Including the positive one held in this study. It is necessary to find out what series of elements need to be put together to help managers improve their businesses. The required empirical work can make use of conceptual guidance to get a reasonably sufficient set of business areas to investigate, like the categories and drivers introduced in this synthesis. The integrated evidence suggests that an effective combination of constructs could yield more useful results in the empirical analysis of SMEs. As mentioned earlier, SMEs face a wide number of challenges. At the same time, they cannot focus only on a few of them just to avoid disruption.

CONCLUSION
The main purpose of this literature synthesis is to provide an alternative framework for the academic and professional assessment of SME performance. This study integrated the evidence from the empirical literature of contingency models, resulting in a categorized set of variables that can explain the positive facet of small firms' outcomes. This paper identifies seven sections from the level of analysis, 24 categories from the disciplines of the original studies, and 131 supported variables that represent a comprehensive framework for the academic and practical scrutiny of the performance of SMEs. Those conceptual findings and integrated elements can help improve the analysis of small firms' performance and explain their meliorated outcome due to their ample, diverse, and global characteristics, which have theoretical pertinence from the original underpinnings and have a proven positive association with benign results in a number of known studies from around the world. This view complements the dominant paradigm of performance models, which have limitations like covariate selection and is overly focused on predicting failure.
This synthesis contributes to the literature of SME performance with an integrative perspective of the previously unconnected empirical literature contained in contingency models. The level of analysis, sections, categories, and drivers developed in this conceptual integration are intended to work as an alternative framework for the academic analysis of SMEs' economic outcomes, acting as classification points and proposed empirical constructs. Those elements can also be employed by professionals as operational insights to improve their firms' results. Their wider and more complete theoretical perspective can help overcome the limitations of performance models by supplying a reasonably ample framework to cover the number of angles in modern business.
It is necessary to continue the academic efforts with both conceptual and empirical research. There is a need for conceptual research to determine the methodologies to be employed in the arena of SME performance to provide consensus on the adequate measures of performance as a criterion, including the convergence of economic and financial units. More integrative work is also necessary to confirm whether the positive perspective and integration suggested here can be generalized. Finally, further empirical research, regardless of the discipline of the precursors, is needed to establish the right combination of categories and variables for a more holistic approach to match the challenging environment faced by SMEs.