Gender and age-based investor affinities in a Ponzi scheme

The study investigates how gender- and age-based affinities between referrers and investors shape the diffusion of a Ponzi scheme and the financial outcomes of participants. Ponzi schemes persist across contexts and often exploit trust within social networks and shared group identities. Despite widespread attention, investor-level analyses with explicit referrer–investor links are rare. Using detailed police records from a 2016 Chinese fundraising Ponzi scheme, the authors aim to open the black box of scheme diffusion by testing whether gender and age affinities: (H1) facilitate scheme spread; (H2) place investors at lower layers in the scheme’s hierarchy; and (H3) reduce investor returns. The work is motivated by evidence that such schemes in China proliferated during 2015–2017 amid mobile/social media, and by policy relevance for protecting vulnerable groups. The contribution lies in documenting investor-level affinity patterns and linking them to hierarchical positioning and losses.

Prior research highlights the role of social networks and trust in propagating financial ideas and fraud. Media and regulators (e.g., U.S. SEC on affinity fraud) note targeting of identifiable groups (ethnicity, religion, age, gender, profession). Academic work shows social interactions spread investment behavior (Stein, 2008; Banerjee & Fudenberg, 2004; Shive, 2010) and models diffusion of Ponzi schemes in networks (Zhu et al., 2017). Studies of Ponzi victims (Smith, 2010; Deason et al., 2015; Gurun et al., 2018) generally use scheme-level data and cannot map referrer–investor links or affinities directly. Rantala (2019) documents referrer–investor links in a Finnish scheme and effects of inviter characteristics (age, education, income) on amounts invested but does not pair on common traits to establish affinity per se and finds no gender effect. In China, descriptive reports suggest women and older adults are disproportionately victimized; Fei et al. (2020) find women and older individuals more likely to participate in a scam but without referrer–investor links. Broader literature shows gender/age differences in risk-taking and financial behavior (Barber & Odean, 2001; Sundaram & Yermack, 2007; Yim, 2013; Serfling, 2014; Faccio et al., 2016), and lower financial literacy among women and older adults (Lusardi & Mitchell, 2008, 2014; Lusardi et al., 2014). The study leverages these insights and the availability of investor-level IDs to focus on gender and age as salient, observable sources of affinity.

Setting and data: A fundraising Ponzi scheme operated via shell company X in a major Chinese city from March to August 2016. Investors purchased RMB 3,900 units with a promised nine-week stream totaling RMB 7,400; principal and interest could be reinvested. The scheme paid referrers and exhibited a pyramidal recruitment structure. Police raided the firm in August 2016 and obtained trading and bank records. Dataset: 4,843 investors from 29 provinces; total raised ≈ RMB 260 million; total loss ≈ RMB 198 million. Investor-level data include bank transactions and national ID-derived gender, age (in 2016), and birthplace. The dataset provides hierarchical referrer–investor links and investor layer (0–22). Sample covers the whole scheme period. Variables: Female_investor (indicator), Age_investor (years), Poverty_investor (born in state-designated poor county), Famine_investor (born 1947–1961, proxy Great Famine youth experience). For referrers: Female_referrer, Age_referrer, Poverty_referrer, Famine_referrer. Outcome variables: Interval (days from 2016-03-05 to joining), Invest_Freq (number of investments per investor over scheme life), Layer (higher value = lower layer), Return = (withdrawals − investments)/investments. Investor participation timing and activeness: Cox hazard model for time-to-entry: h(t) = h0(t) * exp[α1 Female_investor + α2 Age_investor + α3 Poverty_investor + α4 Famine_investor]. Investment activeness: OLS and Poisson regressions of Invest_Freq on Female_investor, Age_investor, Poverty_investor, Famine_investor. Spread via affinities: Using referrer–investor links, estimate (i) Logit: prob(Female_investor=1) on Female_referrer, Age_referrer, Poverty_referrer, Famine_referrer; (ii) OLS and Poisson: Age_investor on Female_referrer, Age_referrer, Poverty_referrer, Famine_referrer. Tests H1. Investor hierarchy: OLS and Poisson of Layer_investor on (i) investor gender/age controls; (ii) affinity pair indicators: Male-to-Female, Female-to-Male, Female-to-Female; Young-to-Old, Old-to-Young, Old-to-Old; plus investor and referrer controls. Tests H2. Investor performance: OLS of Return_investor on (i) investor gender/age controls; (ii) affinity pair indicators and investor/referrer controls. Tests H3. Additionally, regress Return on Layer to link hierarchy to losses. Controls and rationale: Poverty and Famine capture early-life hardship memories potentially affecting behavior (per Bernile et al., 2017; Xu & Li, 2016). Robust standard errors used for OLS; Poisson for counts; odds ratios reported for logit.

Descriptive: 60.89% of participants are female (mean Female=0.6089). Mean age 43.38 years (median 44). Average layer 5.95 (0–22). Average investment frequency 23.38 (median 10). Mean Return −0.4906 (median −0.5860), implying average loss ≈ 49%. Participation timing and activeness: Cox model shows no significant effect of Female or Age on time of entry (Female 0.0352, z=1.14; Age 0.0007, z=0.41). Female and older investors are more active: Invest_Freq OLS Female 2.4890 (t=1.74), Poisson Female 0.1097 (z=1.75); Age OLS 0.2603 (t=3.82), Poisson 0.0118 (z=4.02). Spread via affinities (H1): Female investors are more likely to be referred by female referrers (Femalereferrer 0.2355, z=3.81; odds ratio 1.2656). Femalereferrer age effect also positive (Agereferrer 0.0127, z=3.44). Older investors are more likely to be referred by older referrers: Age_investor on Age_referrer OLS 0.3219 (t=15.06), Poisson 0.0076 (z=14.99). Older investors are also more likely to be referred by female referrers (Femalereferrer 1.1616, t=3.67; Poisson 0.0270, z=3.66). Cross-affinity is observed (female↔older). Hierarchy placement (H2): Female investors occupy lower layers (Layer OLS Female 0.3027, t=3.41; Poisson 0.0514, z=3.40). Older investors also at lower layers (Age 0.0196, t=3.97; Poisson 0.0033, z=3.98). Affinity pairs: Female-to-Female positive and significant (OLS 0.3931–0.4138; t≈3.08–3.28; Poisson 0.0658–0.0689, z≈3.04–3.21). Old-to-Old positive and significant (OLS 0.4498–0.4538; t≈3.41–3.44; Poisson 0.0744–0.0750, z≈3.45–3.48). Other pairings largely insignificant. Investor performance (H3): Female and older investors have lower returns (Female −0.0285, t=−1.82; Age −0.0018, t=−2.01). Affinity pairs reduce returns: Female-to-Female −0.0505 (t=−2.14) and −0.0502 (t=−2.15); Old-to-Old −0.0573 (t=−2.71) and −0.0571 (t=−2.71). Lower hierarchical position is associated with larger losses: Return on Layer coefficient −0.0204 (t=−9.05, untabulated). Collectively, gender/age affinities increase susceptibility, push investors to lower layers, and exacerbate losses.

Findings directly support the hypotheses. H1: Gender- and age-based affinities facilitate scheme diffusion; investors are more likely to be recruited by referrers sharing the same gender or higher age, with additional cross-affinity (female investors referred by older referrers and older investors referred by female referrers). H2: These affinities place investors at lower hierarchical layers, consistent with late entry and weaker payoff prospects in a pyramid. H3: Affinity-linked recruitment reduces returns by roughly 5–6 percentage points in female-to-female and old-to-old pairs, beyond baseline losses, and lower layers predict more negative returns. The results suggest that trust and empathy within demographic groups, potentially combined with lower financial literacy and social-media-enabled networks, make female and older investors particularly vulnerable in affinity-based recruitment. The study advances the literature by empirically establishing investor-level affinity links and showing their consequences for hierarchical positioning and losses, offering practical implications for targeted investor protection and enforcement.

The paper shows that in a Chinese Ponzi scheme, gender and age-based affinities between referrers and investors are pronounced and consequential. Female and older investors are more likely to be recruited by female or older referrers, occupy lower layers in the scheme’s hierarchy, and experience worse returns, with female-to-female and old-to-old links especially detrimental. These investor-level insights contribute to understanding how Ponzi schemes exploit social trust and affinity. Policy implication: regulators and law enforcement should prioritize education and outreach to female and older populations and monitor affinity-based recruitment channels, particularly on mobile/social media. Future research should access broader multi-scheme datasets, include non-participants to model entry probabilities, and examine interactions with broader market conditions (e.g., stock and real estate markets).

The study analyzes a single Ponzi scheme, raising concerns about representativeness and external validity given thousands of contemporaneous cases in China. Data constraints prevent assessing the likelihood of initial participation among the general population; only participants are observed, so entry decisions among those approached but refusing are unobserved. Some variables (poverty and famine) proxy early-life experiences and may not capture individual-level hardship precisely. Proprietary police data are not publicly available, limiting replication. Results may be specific to the 2016 mobile/social-media context and the scheme’s design.