Multi-Level Marketing Participation and Social Connectivity

A TFI Research Challenge project by Claes Bäckman & Tobin Hanspal
Posted on March 19, 2018

The research project "Multi-Level Marketing Participation and Social Connectivity" is one of the eight projects selected last year for the TFI Research Challenge. In this report, the authors use data on social network connectivity and demographic characteristics to identify the determinants of Multi-Level Marketing activity.


Over the past decade the share of employees working as independent contractors has increased dramatically. While this trend has been prominently discussed through the lens of the online “gig”-economy,” much of this work is facilitated via offline firms (Katz and Krueger, 2016). In particular, one aspect of the rise of independent contractors that has received limited attention is direct selling businesses, where individuals work independently and sell goods provided by a parent company. More than five million Americans had a part-time or full-time involvement as a business owner running a direct-selling business in 2016, corresponding to approximately three percent of the labor force. In total, the number of individuals directly affiliated with such firms grew by 30% between 2011 and 2016, from 15.6 million in 2011 to 20.5 million in 2016 (Direct Selling Association, 2017a). Considering the relatively large share of the labor force involved in direct selling and its growing importance, the attention on the direct-selling industry has been surprisingly limited.

Even as the industry is growing in importance, anecdotal evidence suggests that involvement with a direct-selling intermediary can end in financial losses. This is particularly relevant for so-called Multi-Level Marketing (MLM) firms, where the business model relies on individual retailers' commission-based product sales direct to customers and on the ability to recruit new members into the company. Most individuals who join an MLM do not experience any financial gains, in contrast to the often rosy marketing claims (Taylor, 2011; FTC, 2016a). We focus specifically on MLM businesses in an initial attempt to improve our understanding of why individuals invest in projects that regulators continue to warn consumers against. As state, federal, and non-governmental organizations paint an increasingly precautionary view of the MLM industry, understanding the economic determinants and consequences of joining a MLM is becoming more important. In fact, the Federal Trade Commission (FTC) in the US warns: “Not all multi-level marketing plans are legitimate. If the money you make is based on the number of people you recruit and your sales to them…It could be a pyramid scheme.”

The overall social network connectivity of a county is an important correlate of MLM participation in the US.

Our study provides an initial examination of direct-selling firms by focusing on understanding the drivers behind the choice to join a MLM. We obtained data from a Freedom of Information act to the FTC regarding individuals who received compensation from a large MLM company. The FTC investigated and filed a lawsuit claiming that the company made misleading moneymaking claims and that it incentivized its distributors to recruit other members, rather than selling its own product – a violation of the FTC act designed to combat Ponzi schemes. The company's settlement was used to refund over $200 million dollars to nearly 350,000 independent individual distributors exposed between 2009 and 2015. Our data gives us geographic information on each one of the distributors as well as their personal settlement check – a rough proxy for the size of their losses. We aggregated the individual-level data to the county-level and investigated the link between the incidence of MLM activity in a county and social network connectivity, income, demographic characteristics, and social capital, providing new insights into an increasingly important phenomenon in today's labor markets. One of our primary hypotheses was that participation in a multi-level marketing business is driven partly by social behavior. A MLM business by definition relies upon individual members continuously recruiting new members, often through networks of friends or family. The Direct Selling Association states that more than 70% of sales of all MLM businesses are through a direct person-to-person channel, while an additional 20% are facilitated by ‘group sales’ (Direct Selling Association, 2017a). Both of these channels increasingly rely on online social networks, where indeed a rising share of MLM sales are conducted. The New York Times noted in a recent article that while group or event-sales have existed since at least the 1940s, more modern distributors “... add a contemporary spin with the use of e-commerce, mobile credit card swipers, and heavy use of Facebook, YouTube and Twitter” (Dunn, 2016). Our results suggest that the overall social network connectivity of a county is an important correlate of MLM participation. A county with more connections may have larger opportunities to profit from a MLM business, as the potential for retail sales and recruitment is larger. Moreover, we find that counties with MLM incidence are connected to other counties with high MLM incidence.

The implication of social participation in MLM activity, and the associated negative experience, is a two-way street. On the one hand, vulnerable households may be more easily recruited into participation if they are specifically targeted by individuals within their network. On the other hand, awareness of the pitfalls and important educational and financial literacy programs designed to combat predatory business opportunities may be spread via social networks. If these types of opportunities are spread via social networks and social connections, the same networks could also be used to help vulnerable households avoid these opportunities likely to cause them harm.

We find that MLM incidence is particularly concentrated in middle-income counties, and in counties with higher income inequality. In addition, we find that MLM incidence is higher in counties with higher female labor force non-participation relative to male non-participation. In other words, if there are more women remaining outside of the labor market relative to men outside of the labor market, MLM incidence is higher. This suggest that secondary earners, often female, may be attracted to a MLM business model because of the increased flexibility that this type of business offers.

Our preferred explanation for our finding that MLM activity is higher in middle-income counties is that joining a MLM requires a certain level of financial resources, which implies that lower-income counties are less able to participate. However, it is also possible that the way that the FTC distributed reimbursements unintentionally excluded low-income counties, as the minimum losses for receiving a check was $1,000. 

MLM is not a substitute for people's initial income but acts as a stream of secondary income.

The results also provide some support for the industry claims that direct selling acts as a source of secondary income. In particular, the higher MLM incidence in areas characterized by higher female non-participation in the labor market echoes claims that women otherwise not looking for a job may find running a MLM business an attractive opportunity. We also find that counties with higher entrepreneurial activity have higher MLM incidence, but not counties with more sole proprietors. This suggest that MLM activity is not a substitute for other types of entrepreneurship. We also do not find evidence that unemployed individuals joined an MLM as a substitute for a traditional job.

Reassuringly, we find that the demographics of our sample seem to correspond to the statistics reported by the Direct Selling Association. In particular, we find that our measure of MLM incidence is correlated with county-level Hispanic share, counties with a larger female share, a larger fraction of females not participating in the labor force relative to men, and a younger age structure. The Direct Selling Association (2017a) reports that 22% of the individuals involved in direct selling were Hispanic compared to their 18%share of the US population, that 74% of individuals involved in direct selling were females, and the age distribution of direct selling skews towards younger individuals. To the extent that we are able to measure it, these results suggest that our results generalize to other firms in the industry. However, given the nature of our data we wish to clearly state that we do not mean to suggest that all MLM engage in questionable behavior.

The above results do not mean that the losses experienced by an individual through an MLM are trivial. In fact, we find that counties with higher income experienced larger losses from joining an MLM, as proxied by the size of the settlement check. We investigate where individuals lost the most by examining the size of their refunds, and find that investment losses were more severe in counties with a higher share of Hispanics and women, women outside of the labor force relative to men, and counties with high income inequality and lower educational achievement.

The same social networks that spread harmful investment opportunities to consumers could instead be used to warn against such investments.

The findings help shed light upon the characteristics which correlate with the uptake in alternative working arrangements and may be specifically important for households that seem to be vulnerable to the downsides associated with these opportunities. We provide directional insights for policy makers and practitioners aiming to reduce the potential harm associated with new employment opportunities.  The same social networks that spread investment opportunities that proved harmful to consumers could instead be used to warn against such investments, helping to promote knowledge about how to better invest in relatively safer opportunities in order to achieve financial security.

Dr Claes Bäckman is a Postdoctoral Researcher at the Department of Economics at Lund University and Knut Wicksell Centre for Financial Studies, Sweden. E-mail:

Dr Tobin Hanspal is an Assistant Professor at Research Center SAFE, Goethe University Frankfurt, Germany.  E-mail: