1.  Introduction

Household consumption constitutes a fundamental driving force behind China’s economic growth. Effectively stimulating residents’ consumption potential and optimizing the consumption structure have emerged as pivotal issues within the current strategic framework aimed at promoting high-quality development and advancing the domestic demand. Informal institutions—such as collectivist values, clan-based culture, and neighborhood ties—play a critical role in shaping economic behavior in China. By strengthening social capital, facilitating social interactions, and promoting information dissemination, informal institutions enhance the efficiency of household financial decision-making, thereby fostering both the elevation of consumption levels and the upgrading of consumption patterns. In recent years, against the backdrop of profound socio-economic transformation, the explanatory power of traditional “hard” constraints—such as income and prices—on consumption behavior has gradually declined, while “soft” constraints like social interaction have become increasingly prominent. China’s historically entrenched acquaintance society, gift-giving traditions, and neighborhood networks form vital channels for information transmission and behavioral imitation among households, representing significant dimensions for elucidating heterogeneity in household consumption. As a core component of informal institutions, social interaction operates through mechanisms such as information sharing, behavioral mimicry, and social norms, reshaping households’ financial cognition and risk preferences, and consequently influencing consumption decisions [1]. Since humans are inherently social beings, their financial and consumption behaviors are inevitably shaped by their social networks. The notion of consumption upgrading encompasses not only structural optimization but also a sustained increase in total consumption volume [2]. The mechanisms through which social interaction contributes to this upgrading merit further scholarly investigation.

Although existing studies have highlighted the positive impact of social interaction on household consumption, the current literature suffers from insufficient identification of underlying mechanisms and fragmented analytical perspectives. In particular, there is a lack of systematic research on the mediating effects through which social interaction influences consumption upgrading via financial literacy accumulation and risk-sharing pathways. In response, this paper from a micro-level household perspective, focusing on the dual-path mechanism of “literacy and risk” to systematically examine how social interaction promotes household consumption upgrading by optimizing financial decision-making. A mediation model is constructed to identify the transmission roles of financial literacy and risk management in the relationship between social interaction and consumption behavior and investigates heterogeneity across urban-rural areas and regional dimensions. This research contributes to a deeper understanding of the relationship between informal institutions and household consumption behavior, providing theoretical support and policy insights for enhancing social capital, promoting financial literacy, and advancing high-quality development in household consumption.

The remainder of this paper is organized as follows: Section 2 outlines the theoretical framework; Section 3 presents the research design; Section 4 reports and analyzes the empirical findings; Section 5 conducts the mechanism analysis; and Section 6 concludes with policy implications.

2.  Theoretical hypotheses

Extensive research has confirmed that social interaction is a key factor influencing household consumption upgrading. It primarily operates through mechanisms such as information sharing, behavioral demonstration, and trust building, effectively facilitating the transformation of household consumption from subsistence-oriented to development- and hedonic-oriented patterns. Existing literature has mainly explored this influence from two dimensions: consumption structure optimization and consumption expenditure growth. Regarding consumption structure, Qin Xinran’s study based on data from the China Family Panel Studies (CFPS) found that social interaction has a particularly significant effect on the enhancement of development-oriented consumption—such as education and health—and hedonic consumption, including leisure and tourism [3]. Specifically, social interaction promotes consumption structure optimization by raising households’ awareness of and preferences for high-quality consumption. In terms of consumption level, social interaction directly increases household consumption expenditure by reducing information asymmetry and boosting consumer confidence. He Yuan et al., focusing on the elderly population, found that social activities like group dancing drive consumption structure upgrading through dual pathways of “comparison effect” and “cognitive effect.” Long-term and frequent social interaction can raise household consumption expenditure, thereby empirically validating the impact of social interaction on consumption from the perspective of total expenditure growth [4]. Additionally, the peer effects within social networks can stimulate household consumption willingness and significantly improve acceptance of emerging products and services such as smart home technologies and health services. Based on the above analysis, this paper proposes the following hypothesis:

Hypothesis 1: Social interaction positively promotes household consumption, including increases in total household consumption expenditure and the optimization of consumption structure.

A substantial body of literature shows that the impact of social interaction on household consumption upgrading is not homogeneous but exhibits significant urban-rural and regional heterogeneity. Generally, due to broader information channels and more consumer choices, the role of social interaction in promoting consumption upgrading is more pronounced among urban households and those in economically developed eastern regions. Zhou Lan revealed that internet-driven social interaction significantly enhances offline consumption in urban households, but its effect is limited for rural households due to inadequate infrastructure, confirming the “digital divide” as a constraint on consumption upgrading [5]. The eastern region, characterized by economic development and mature financial markets, shows a more significant promoting effect of social interaction on consumption upgrading. Wu Yufeng et al. further revealed notable spatial heterogeneity in the effect of social interaction on insurance participation, particularly among urban and high-income residents in the east [6]. Financial literacy has a greater impact on increasing survival consumption levels among rural households and those in central and western regions, while it has a larger effect on consumption structure optimization in urban households and eastern regions [7]. Based on the above literature, the following hypothesis is proposed:

Hypothesis 2: The effect of social interaction on household consumption upgrading exhibits significant urban-rural and regional heterogeneity, with a more pronounced impact on households in urban and eastern regions.

Social interaction influences household consumption upgrading primarily through two core channels: financial literacy accumulation and risk-sharing. First, in terms of financial literacy accumulation, social interaction accelerates the dissemination and spillover of financial knowledge through interpersonal communication, effectively reducing households’ information acquisition costs and thereby enhancing their financial literacy [8]. This effect is rooted in the “neighborhood effect” emphasized by Durlauf, [9] whereby individuals optimize their financial decisions through endogenous interactions with others, such as information exchange and observational learning [10]. Wang Lin found that social interaction significantly improves the effectiveness of rural households’ financial asset portfolios, with financial literacy playing a dominant mediating role [11]. Improved financial literacy facilitates a shift from savings- oriented to diversified investments strategies, [12] increasing the proportion of risky assets held, which releases a wealth effect [7] and subsequently promotes a higher share of development-oriented consumption. Social interaction, social trust, risk preferences, and family responsibility jointly influence household consumption levels. Social interaction promotes the optimization of household financial decisions through demonstration effects and peer pressure, [13] and mitigates the consumption- inhibiting effect of future uncertainty via commercial insurance participation. Commercial insurance reduces precautionary savings for unexpected expenses through risk transfer [14]. Based on the above literature, the following hypothesis is proposed:

Hypothesis 3: Financial literacy accumulation and risk-sharing play significant mediating roles in the effect of social interaction on household consumption upgrading.

3.  Model specification and data sources

3.1.  Model specification

This paper adopts the Ordinary Least Squares (OLS) regression as the baseline model, specified as follows:

$$$ Y=\alpha +\beta \times SocialInteraction+\gamma \times Controls+\epsilon $$$

Where:  $$ Y$$  represents the proxy variable for household consumption level, including the logarithm of total household consumption expenditure and the proportion of developmental/hedonic consumption in total consumption;  $$ SocialInteraction$$  denotes social interaction, with  $$ \beta $$  as the corresponding regression coefficient representing the marginal effect of social interaction on household consumption;  $$ Controls$$  is a vector of other control variables;  $$ \epsilon $$  is the random error term.

To examine the mediating mechanisms of financial literacy accumulation and risk-sharing in household consumption upgrading, this study applies a two-step approach to test the mediation effects of financial decision optimization. The models are specified as follows:

$$$ {Intern}_i={\sigma }_i+{\theta }_i\times SocialInteraction+{\phi }_i\times Controls+{\mu }_i$$$

$$$ Y={\alpha }_i+{\beta }_i\times SocialInteraction+{\beta }_i\times {Intern}_i+{\gamma }_i\times Controls+{\epsilon }_i$$$

First, we test whether social interaction alters household financial decisions ( $$ {Intern}_i$$ ). Second, we incorporate indicators of financial decision optimization as control variables into the baseline regression to test whether these indicators positively influence household consumption and whether their inclusion attenuates the effect of social interaction on consumption: If  $$ {\theta }_i$$  is significantly positive,  $$ {\beta }_i$$  is also significantly positive, and  $$ {\beta }_i$$  <  $$ \beta $$  with a reduced significance level, it indicates that financial literacy accumulation and risk-sharing serve as effective mediators influencing household consumption.

3.2.  Variable description

3.2.1.  Dependent variables

Household consumption is analyzed from two perspectives: the growth of total household consumption expenditure and the optimization of consumption structure. The total consumption expenditure is measured as the logarithm of all consumption expenditures incurred by the household, including both survival-oriented and hedonic consumption. The consumption structure is captured by the proportion of hedonic consumption within total consumption expenditure. Friedrich Engels posited that human consumption needs can be categorized into survival, developmental, and hedonic. After basic survival needs are met, individual consumption preferences gradually shift towards personal improvement and spiritual fulfillment. Based on the classification standards commonly used National Bureau of Statistics and household survey data, this paper classifies household food consumption, utilities (including water, electricity, fuel), heating costs, property management fees, daily necessities, and clothing as survival consumption; while spending on beauty and bathing services, personal care products, medical and healthcare supplies, domestic services, housing renovation and maintenance, transportation and communication, cultural and recreational activities, education and training, tourism, and online shopping are classified as hedonic consumption. These reflect consumption aimed at better quality of life and spiritual satisfaction.

3.2.2.  Core independent variables

Social Interaction is measured following the methodology of Du et al., by selecting seven expenditure categories related to social interaction: monetary gifts, dining out, entertainment, communication, transportation, tourism, and visiting relatives [15]. These variables are standardized as proportions of total household income and subjected to factor analysis using the iterative principal factor analysis to construct a composite index that comprehensively reflects the degree of social interaction.

LR test: independent vs. saturated: chi2(21) = 1.7e+05 Prob>chi2 = 0.0000

To assess the suitability of the seven selected variables for factor analysis, the Kaiser-Meyer-Olkin (KMO) test was applied. According to the KMO principle, values closer to 1 indicate that the original variables share common factors and are thus appropriate for factor analysis, whereas values closer to 0 indicate weak correlations unsuitable for factor analysis. The overall KMO value of the seven variables selected is 0.652, which exceeds the threshold of 0.5, indicating adequacy for factor analysis. Principal component analysis results show that two components have eigenvalues greater than 1. Therefore, the first two factors—monetary gifts and visiting relatives expenditure—were selected to measure social interaction, with a cumulative contribution rate approaching 80%. Finally, the composite social interaction score was calculated as a weighted sum of the component scores, where weights correspond to each component’s contribution ratio to the cumulative variance.

3.2.3.  Mediating variables

Financial Literacy: Generally, households with greater attention to economic and financial information possess higher financial literacy levels. Therefore, this paper uses the degree of household attention to economic and financial news as a proxy for financial literacy to test the mediating effect of social interaction on household consumption upgrading. A dummy variable was created based on the survey question, “How much attention do you usually pay to economic and financial information?” Respondents’ answers ranged from “never” to “very often,” coded as 1 to 5 respectively, with higher values indicating greater attention.

Risk Management: The paper assesses household risk management by the participation in commercial insurance. A household is considered to have purchased commercial insurance if at least one family member holds any type of commercial insurance. A dummy variable for household insurance purchase is set as 1 for insured households and 0 otherwise.

3.2.4.  Control variables

Following the approach of Li Xiang et al., this paper controls for characteristics at both the household head and household levels [7].

At the household head level, characteristics include gender, age, education level, ethnicity, health status, and marital status. Specifically, gender is coded as 1 for male and 0 for female; age is calculated as the difference between the survey year and the respondent’s birth year; education level is represented by a dummy variable, taking the value 1 if the household head has a bachelor’s degree or above, and 0 otherwise; ethnicity is a dummy variable coded as 1 for ethnic minorities and 0 for Han ethnicity; health status is based on the survey question, “Compared to people of the same age, how is your current health status?” with “healthy” or better coded as 1 and others as 0; marital status is a dummy variable based on the question, “What is your current marital status?” with “married” or “cohabiting” coded as 1 and others as 0.

At the household level, characteristics include total household income, total household assets, household size, elderly dependency ratio, child dependency ratio, and housing ownership. Total household income and total household assets are measured by their natural logarithms; household size is measured by the number of household members; elderly dependency ratio is the proportion of household members aged 65 and above; while the child dependency ratio is the proportion of household members aged under 14. Housing ownership is measured by the number of properties owned by the household, based on the survey question, “Excluding rented houses, how many houses does your household own?”

3.3.  Data source

This paper utilizes data from the 2019 China Household Finance Survey (CHFS), a nationally representative dataset that provides a comprehensive overview of the financial conditions and economic behaviors of urban and rural households in China. The CHFS dataset contains detailed information on household financial assets, income, insurance participation, social relationships, and other aspects, making it well-suited for empirical analysis of micro-level financial behaviors.

4.  Empirical results and analysis

4.1.  Baseline regression results

The baseline regression results are presented in Table 2. Columns (1) and (4) show the effects of the core explanatory variable—social interaction—on household consumption upgrading when controlling for individual-level characteristics of the household head. Columns (2) and (5) include only household-level control variables, while Columns (3) and (6) incorporate all control variables. The core findings indicate that even after controlling for all variables, the promoting effect of social interaction on household consumption remains statistically significant. Specifically, Column (3) shows that a one-unit increase in the level of social interaction is associated with a 6.9% increase in total household consumption expenditure, while Column (6) shows a 0.6 percentage point rise in the share of hedonic consumption. Both effects are statistically significant at the 1% level. These findings support the central hypothesis of this paper that social interaction is an important driver of household consumption upgrading.

Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01

4.2.  Heterogeneity analysis

Considering China’s dual urban-rural structure and regional development imbalances, this paper further examines whether the effect of social interaction on household consumption exhibits heterogeneity across urban-rural and regional dimensions. To this end, the full sample is divided into urban and rural, as well as eastern and central-western subsamples for separate regressions. The results are shown in Table 3. Comparison of Columns (1) and (2) indicates that the boosting effect of social interaction on household consumption is more pronounced among rural households. Meanwhile, Columns (3) and (4) reveal that the promoting effect of social interaction on household consumption is stronger in the eastern region compared to the central and western regions.

Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01

4.3.  Robustness checks

4.3.1.  Replacing the core dependent variable

Given the inherent differences among types of consumption, survival consumption is constrained by rigid factors such as income, while hedonic consumption better reflects proactive consumption decisions induced by social interaction. Therefore, hedonic consumption expenditure is used as an alternative proxy for household consumption upgrading to conduct a robustness check. Columns (1) to (3) of Table 4 report regression results using hedonic consumption expenditure as the dependent variable. The results show that the positive impact of social interaction on household consumption remains robust, with coefficients significantly positive at the 1% level, further confirming the reliability of the main findings.

4.3.2.  Replacing the core independent variable

To further verify model robustness, this paper replaces the core explanatory variable. In the Chinese social context, expenditures on social monetary gifts represent an important cost for maintaining social networks and serve as an effective quantitative proxy for household social interaction. Drawing on related studies, [4] monetary gift expenditure is used as an alternative proxy variable for social interaction in the regressions. Columns (4) to (6) of Table 4 show that, after substituting the proxy, the main conclusions remain robust. Monetary gift expenditure has a significant positive effect on both total household consumption expenditure and consumption structure optimization at the 1% level, further validating the robustness of the baseline regression results.

Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01

4.4.  Endogeneity test

To address potential endogeneity issues arising from omitted variables or reverse causality, this paper employs an instrumental variable (IV) approach. Following established practice, [8] a community-level social interaction index is selected as the instrument for individual-level social interaction. Specifically, to mitigate the influence of outliers, the median social interaction level of households within the community is used to measure the overall intensity of community interaction. This instrument satisfies the two core validity conditions: first, the relevance condition is met. The social atmosphere within a community shapes individual social habits, and the two are significantly correlated. The F-statistic from the first-stage regression is 40 (well above the threshold of 10), with a p-value less than 0.01, rejecting the null hypothesis of a weak instrument. Second, the exclusion restriction reasonably satisfied, as the overall community interaction level is determined by factors exogenous to individual household decisions (such as community customs) and does not directly affect the household consumption decisions. The IV regression results reported in Table 5 confirm that the core conclusions remain robust, demonstrating that the baseline regression results are not driven by endogeneity.

Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01

5.  Mechanism analysis

To examine the mediating roles of financial literacy accumulation and risk-sharing, this section conducts a mechanism analysis.

5.1.  Financial literacy

Regression results in Columns (1) to (3) of Table 5 show that social interaction has a significant positive effect on household financial literacy levels. These results indicate that social interaction effectively reduces the cost of acquiring financial information through interpersonal communication and knowledge spillover, thereby improving their financial literacy. Improved financial literacy contributes to more efficient asset allocation and investment decision-making, enhances property income, and ultimately provides a solid foundation for consumption upgrading. These results provide empirical support for the mediating role of financial literacy proposed in Hypothesis 3.

5.2.  Risk sharing

Columns (4) and (5) of Table 6 test the risk-sharing mechanism. The results reveal that social interaction significantly increases the probability of households purchasing commercial insurance. Commercial insurance, as an effective risk management tool, helps households hedge against future uncertainties and reduces their motivation for precautionary savings. The liquidity released from lower precautionary savings can then be used to increase current consumption levels and optimize the consumption structure. Therefore, the mediating role of risk sharing proposed in Hypothesis 3 is also empirically supported.

Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01

6.  Conclusion

Social interaction significantly increases both the total household consumption expenditure and the proportion of hedonic consumption. By promoting the accumulation of financial literacy and risk-sharing behaviors, social interaction further optimizes household financial allocation, strengthens consumption confidence and expectation stability, thereby driving the transformation of household consumption from survival-oriented to development- and hedonic-oriented structure. This effect exhibits heterogeneity across urban-rural and regional dimensions, with a more pronounced promoting effect on households in the eastern region and rural areas, highlighting how regional development disparities and infrastructure conditions shape the pathways to consumption upgrading.

The findings of this study yield important practical implications for continuously advancing high-quality household consumption and unleashing the potential of domestic demand. The specific policy recommendations are as follows: First, the positive guiding role of micro-level social networks — such as communities and social groups — should be fully leveraged by cultivating a proactive and interactive community environment that facilitates information sharing and financial knowledge dissemination among households, thereby stimulating residents’ willingness to consume; Second, multi-level and multi-channel financial literacy initiatives should be promoted , with particular attention given to households in the central-western and rural areas, to narrow the financial literacy gap and enhance their wealth management capabilities; Third, the supply of commercial insurance systems should be improved and the coverage of inclusive risk management tools should be expanded, guiding residents to actively participate through social interaction mechanisms, thereby releasing precautionary savings and boosting consumption vitality.

Social interaction, as an essential component of informal institutions, not only influences household behavior but also plays a critical role in consumption upgrading by optimizing financial decisions and mitigating economic uncertainty. Future research should further explore micro-level household mechanisms to promote the formation of a virtuous cycle linking social interaction, financial optimization, and consumption upgrading.