Gift-giving is a widespread social custom, particularly significant in Chinese culture, often involving the concept of *renqing*. While existing literature explores gift-giving in various contexts and cultures, including its role in signaling care, meeting expectations, and potential misinterpretations (Givi & Galak, 2021; Galak et al., 2016; Minowa & Belk, 2018; Li et al., 2021; Li & Peng, 2021), a comprehensive understanding of the influence of social conformity on gift-giving in China remains lacking. This paper aims to fill this research gap by examining the effect of social norms on gift-giving behavior in China at a national level using a large representative sample. The study grounds its analysis in the law of diminishing marginal utility, which posits that the added satisfaction from giving more gifts decreases as the amount given increases. By incorporating this principle, the study offers a unique perspective on the motivations behind gift-giving decisions, particularly how the actions of others influence these decisions. The study's large and representative dataset, drawn from the China Family Panel Studies (CFPS), enables researchers to generalize findings beyond controlled laboratory or small-scale field experiments. This approach also addresses the limitation of previous studies that primarily focused on laboratory or small-scale field experiments. The research also acknowledges the importance of social belonging and group identification as motivations driving gift-giving behavior, highlighting the prevalence of conformity in shaping these behaviors.

Existing research on gift-giving primarily focuses on material payoff or positive reciprocity (Camerer, 2003; Stanca et al., 2009). However, a significant body of work underscores the impact of social norms and the influence of others’ actions on individual behavior (Francisco et al., 2008; Kuran, 1995). In China, gift-giving, often expressed through the concept of *renqing*, plays a crucial role in social relationships and business interactions (Steidlmeier, 1999; Chan et al., 2003; Yeung & Tung, 1996; Hwang, 1987; Wang, 2007). Studies highlight the reciprocity inherent in *renqing* (Steidlmeier, 1999; Chen, 1995), its role in building trust in business transactions (Steidlmeier, 1999; de Bary, 1991), and its influence on purchase intentions and leadership dynamics (Zhang et al., 2022; Ren et al., 2021; Ren et al., 2020; Yen et al., 2017; Chen et al., 2018). Despite this existing literature, there's a lack of large-scale, nationally representative studies examining gift-giving conformity in China. This study addresses this gap by using a substantial dataset to analyze the prevalence and determinants of following behavior in gift expenditure in China. This addresses the gap in literature by providing a large scale study on gift giving conformity in China. The study's focus on national-level data and the use of an established statistical model distinguish it from previous research.

The study utilizes data from the China Family Panel Studies (CFPS) from 2014 to 2018, encompassing three waves and 37,147 individuals. A key variable, *a_it*, represents an individual's gift-giving expenditure in the past 12 months. A relative index (*a_it*/ *a_it***) is calculated by dividing individual expenditure by the leave-out average expenditure of peers in the same age and community group. This index captures the degree to which an individual conforms to peer gift-giving norms. Statistical models are derived from utility functions, assuming diminishing marginal utility. The study employs several regression models to analyze the data: 1. **Continuous Regression (Equation 2):** This initial regression model examines the relationship between an individual's gift-giving expenditure (*a_it*) and the leave-out average of their peers' expenditures (*a_it***), controlling for other variables. The coefficient *β₁* indicates the effect of peer influence. However, this approach does not estimate conformity at the individual level. 2. **Utility Function (Equation 4):** A utility function *f(a_it, a_it***) is introduced, modeling utility as a function of direct gift-giving effects *u(a_it)*, relative effects *v(a_it/a_it***), and cost *c(a_it)*. This allows for analysis of an individual's tendency to conform (indicated by the sign of ∂a_it/∂a_it***). 3. **Panel Ordinal Discrete Regressions (Equations 7 & 8):** Because life satisfaction (a proxy for utility) is measured as an ordered categorical variable (1-5), panel ordered-outcomes regressions are employed. The authors use both fixed-effects and random-effects ordered logit/probit models, with a dynamic model (equation 8) further accounting for serial correlation and potential endogeneity using lagged dependent variables and Lewbel's instruments. This model addresses potential endogeneity issues by incorporating lagged dependent variables and instrumental variables. 4. **Estimation of Following Behavior:** The sign of ∂a_it/∂a_it*** is determined for each individual using comparative static analysis from the utility function. An indicator variable *w_it* is created (1 if ∂a_it/∂a_it*** > 0), and the percentage of followers is calculated using equations 11 and 12. Bootstrapping is used to obtain standard errors and confidence intervals for these estimates. The study includes various control variables to account for potential confounders in the analysis: age, gender, marital status, urban/rural location, education, social status, family size, income, medical expenses, real estate value, savings, debt, and various expenditure categories. These variables account for various demographic and socio-economic factors impacting gift giving and life satisfaction. Robust regression techniques are applied to handle potential heteroskedasticity and outliers.

The study's key findings are presented across several tables. Table 3 shows that the coefficients for the leave-out average renqing expenditure are significantly positive across different regression models for renqing expenditure as a dependent variable, indicating a positive peer effect on gift-giving behavior. This effect, however, is diminishing, with coefficients less than 1. Additional findings suggest an inverted U-shaped relationship between renqing expenditure and age, with individuals around age 45 having the highest expenditure. Men tend to spend more than women, urban residents more than rural residents, and higher education levels are associated with higher renqing expenditure. Wealthier individuals, as indicated by income and various asset and expense variables, also tend to spend more. Cosmetology expenditures show a particularly strong positive relationship with renqing expenditure. Table 4 presents findings from regressions using utility (measured by life satisfaction) as the dependent variable. The results consistently show a positive relationship between renqing expenditure and utility, with a diminishing marginal utility effect. A U-shaped relationship between age and utility is observed, with young and old people reporting higher satisfaction. Other key findings include higher utility for married individuals, women, urban residents, and those with higher incomes, social status, and cosmetology expenditures. Medical expenses are negatively associated with utility. Table 5 summarizes the estimated percentage of individuals who follow others in renqing expenditure. The results consistently show that over half (approximately 49-61%) of the population follows others, though this proportion decreased from 2014 to 2018. Men are more likely to follow than women, and younger individuals (ages 22-40) exhibit a higher propensity for following behavior than older people. Table 6 shows bootstrapped results, confirming the statistical significance of the estimated following percentages across all models. The confidence intervals across different models partially overlap.

The findings highlight the importance of social conformity in shaping gift-giving behavior in China. The positive and diminishing peer effects suggest that individuals are influenced by their peers but not entirely driven by them. The inverted U-shaped relationship between renqing expenditure and age may reflect life-cycle effects, with individuals at different stages of life having varying financial capabilities and social obligations. The finding of a decreasing trend in conformity from 2014-2018 is noteworthy, possibly reflecting changes in societal norms, government regulations, or increasing individualistic tendencies. The gender and age differences in conformity suggest that social norms exert different influences on various population subgroups. These findings contribute to a nuanced understanding of gift-giving behavior in China, enriching both economic and sociological perspectives. This research also speaks to broader theories of social influence and conformity. The concave relationship between relative renqing expenditure and utility supports the law of diminishing marginal utility. The study's results have implications for marketing, policymaking, and cross-cultural understanding of gift-giving practices.

This study provides the first large-scale, nationally representative analysis of conformity in gift-giving behavior in China. Using a novel methodological approach combining utility functions and panel ordered-outcomes regressions, the study reveals substantial levels of conformity, particularly among younger men. The decline in conformity over time warrants further investigation. Future research could explore the impact of policy changes, evolving cultural norms, and other factors on conformity levels. Expanding the study to other cultures with similar social norms would enhance cross-cultural comparisons and generalize the findings further.

The study's primary limitation is its reliance on self-reported data. Although the CFPS dataset is large and nationally representative, self-reported data on gift-giving expenditure and life satisfaction may be subject to recall bias and response error. The study's focus on China might limit the generalizability of findings to other countries with different cultural contexts. While the authors used advanced statistical techniques, including instrumental variable approaches to tackle potential endogeneity and control for heteroskedasticity, there is always a risk of unobserved confounding variables that could impact the results. The time period covered (2014-2018) may also limit the broader longitudinal analysis.