SMEs are inherently risky organisations reflected in high birth and death rates in many countries. This risk of failure is exacerbated by the financial and operational opacity of SMEs, which hamper efforts by authorities and investors to evaluate and value objectively SMEs. Despite a wide spread view that SMEs are risky there is little evidence of generalizable, empirical estimates SME risk and supporting analyses, covering the relationship between operational characteristics risk. Commencing from a premise that there needs to be a robust framework for quantifying SME risk this research examines, using a Malaysian database, risk measures and their relationship to observable attributes of SMEs. The work is important for a range of stakeholders, including government in relation to sustainable economic growth goals and other areas of public policy. Investors, customers, suppliers, employees and society are unlikely to benefit from business churn where some of it is avoidable. Malaysia, one of the larger economies in Southeast Asia (by GDP) is a multi-cultural country with a diversity of ethnicities. In the 1990s, it, along with various other East Asian countries underwent an economic boom that subsided with the onset of the Asian financial crisis in 1998. It is a constitutional monarchy with members of the royal Malay families assuming the role, in rotation, of Head of State. Given its diverse demographic, Malaysia actively practices an affirmative action scheme designed to give indigenous people (the Bumiputera) better access to finance, education and business opportunities to promote greater economic parity across the different ethnicities living in the country. Islam is the State religion with close to 60% of the population practicing the religion. As such, Islamic finance is widely available and has come to characterise Malaysian fiscal policy as being very conservative and risk-averse in accordance with Islamic principles. SMEs in Malaysia are predominantly family-owned and control is often centralised in the hands of the family, with very few outsiders rising to high managerial positions within these businesses. There is little to no separation between management and shareholders as they are typically the same. This lack of separation creates unique governance issues where the CEO and the Chairman of the board are often the same person. This creates a conflict of interest between the principal (owner) and the agent (manager) as there are weaker checks and balances performed against the CEO if they are the head of the Board of Directors. However, in a family business, because the CEO/Chairman is the head of the family, principal-agent conflict is virtually eliminated as the CEO and the Chairman’s goals align. Despite this, principal-principal conflict still exists through leadership tussles, nepotism and poor consensus decision making. SMEs have contributed significantly to the Malaysian GDP and play a very important role in the economic development of Malaysia. The Companies Commission of Malaysia (Suruhanjaya Syarikat Malaysia - SSM) records SME data, as all Malaysian businesses, regardless of their size are required to furnish the authorities with regular annual reports. The initial dataset consists of an unbalanced panel dataset of 400 individual companies over the years 2005 to 2014. After data cleaning and removal of dormant and insolvent companies, this figure is reduced to 303 individual companies. The current research draws on the information covering financial statements; owner ethnicity, age and gender; business location; shareholding and items stored in the database, providing a rich panel of data for businesses over time. Most empirical risk and return measures, relating to models like Capital Asset Pricing Model (CAPM) and Arbitrage Pricing Theory (APT) originate from research with publicly listed companies and initially in large mature western economies. For SME research, where firms are not listed on stock exchanges, other approaches are necessary. Four models are examined in this thesis, viz. the pure-play (PP) beta, accounting beta and two probability of survival models (1 & 2). These models noted in prior research are tractable within the constraints of data typically available for SMEs and are present in the SSM database. A selection of SME characteristics identified in the literature as connected with SME risk, these include financial indicators such as performance and capital structure; firm characteristics such as firm size, firm age, the firm’s sub-industry and firm’s geographical location; and owner demographics such as owner age, largest share percentage, business owner gender and business owner ethnicity. The analysis consists of two parts. First, a correlation analysis of the risk metrics against the other SME characteristics. The correlation analysis shows that PP beta has the highest level of correlation with not only the other risk measures but also the SME characteristics. The other risk measures have lower cases of significant correlation with the accounting beta having the lowest number of significant correlations with SME characteristics. Second, each risk value is regressed against the SME characteristics using the Dynamic Panel Data Generalised Method of Moments (GMM) regression technique, which effectively captures the panel nature of the data and mitigates the threat of endogeneity. The regression results show that the PP beta measurement has the most significant relationships with the selected SME characteristics, most notably in areas of financial performance and gender of the business owner. The other measurements recorded significant relationships in the same areas as well as capital structure. Interestingly, despite Malaysia’s affirmative action policies, there is no significant relationship between any of the risk measures and business ethnicity. Furthermore, despite a large portion of Malaysian SMEs involved in the service sector, the type of industry SMEs are in do not significantly affect their risk either. The presence of family ownership as represented by concentration of ownership in the largest share percentage is not significantly related to risk, indicating that principal-principal conflict does not adversely affect Malaysian SMEs. Firm size and firm age both are not significantly related to SME risk, providing some hope for smaller, younger businesses trying to convince finance providers to invest in them. Owner age is significantly related to the probability of survival model 1, however it does not significantly relate to any of the other risk measures. The identification of significant relationships between capital structure with SME risk is relevant to lenders and can assist them in pricing their loans to SMEs. The finding that neither firm size nor age are not statistically significantly correlated with SME risk has implications for loan terms and bargaining. The research concludes that PP beta, used in conjunction with a probability of survival model provides the most adequate measure of risk for unlisted SMEs. Combined with the results from the regression, this research challenges some long-standing assumptions held regarding the relationship between SME characteristics and risks in the hopes that the findings will influence policy makers and finance providers to give better financial and development support to SMEs.