We construct a measure of public gambling sentiment using the Baidu index and empirically examine the influence of temporal fluctuations in gambling sentiment on the performance of lottery-type stocks. The empirical findings reveal that heightened gambling sentiment correlates with increased selling volume, short-term order imbalances and liquidity in the lottery-type stock market, while diminishing abnormal returns, including the initial returns of IPOs. We attribute this to investors' gambling mentality stemming from the accumulation of wealth rather than the pursuit of excitement, as they view lotteries and lottery-related stocks as interchangeable investment options.

1 INTRODUCTION

The stock market tends to decline during the FIFA World Cup, experiencing irregular trade volumes, known as the ‘World Cup Curse’ (Drummond, 2023; Fjesme et al., 2023). This event coincides with increased lottery participation and a spike in public gambling sentiment, potentially affecting stock market investment (Zhu et al., 2023). This anomaly relates to the impact of gambling sentiment on the stock market (Chen et al., 2021). Some investors, mirroring lottery players, accept higher odds of small losses for chances of significant gains (Agarwal et al., 2022), showing a preference for ‘lottery-type stocks’ (Barberis & Huang, 2008; Frino et al., 2019). Studying shifts in lottery market sentiment to understand investors' gambling behaviour is therefore intriguing.

Previous studies concentrate predominantly on identifying lottery-type stocks and assessing their market performance. The key contributions of Kumar (2009) and Bali et al. (2017) outline the identification methods. Kumar (2009) suggests employing three distinct indicators – idiosyncratic skewness, volatility and nominal stock price – to discern lottery features of stocks. In contrast, Bali et al. (2017) advocate the maximal daily return over a month as a metric. These indicators have broad applications in subsequent studies (Chen et al., 2021; Liu et al., 2020). Exploring lottery-type stock market performance, numerous researchers observe that investors' skewness preferences often prompt an extra premium for positively skewed stocks, leading to inflated valuations and diminished prospective returns (An et al., 2020; Cheon & Lee, 2018). Specifically, in China's market, a notable ‘gaming premium’ exists, with lottery-type stocks being overvalued and yielding 10%–12% less annually than their non-lottery counterparts (Li et al., 2018). Furthermore, this gambling preference extends to initial public offering (IPO) irregularities because IPO stocks embodying a unique lottery-type category are perceived as lucrative, potentially leading to overpricing (Green & Hwang, 2012).

While prior research acknowledges investors' gambling mentality and the market dynamics of lottery-type stocks, it seldom delves into the underlying causes of this mentality. We posit that a common psychological thread may drive the purchase or trading of lotteries and lottery-type stocks. Certain investors, influenced by psychological and cognitive distortions like the ‘illusion of control’ (Perez & Humphreys, 2013), may show a proclivity towards these investments. Moreover, even knowledgeable investors who are cognisant of dismal expected returns may pursue dual objectives. Initially, the quest for wealth or financial gains from these gaming instruments propels investors to tolerate the risk of negative returns (Kumar et al., 2016), as evidenced by the disproportionate representation of low-income individuals in lottery-type stock investments (Kumar, 2009). Concurrently, the thrill of gambling provides mental exhilaration, prompting numerous individuals to invest for recreational or adventurous reasons (Brown et al., 2018; Gao & Lin, 2015).

The Chinese market provides a distinct landscape. Historically and culturally, the Chinese exhibit a higher propensity for gambling than their Western counterparts, accompanied by a higher prevalence of gambling addiction (Tse et al., 2010). Compounded by the legal prohibition on gambling in China, purchasing or trading lottery or lottery-type stocks emerged as the sole legal avenue for fulfilling gambling inclinations. Notably, individual investors, who constitute a significant segment of the Chinese stock market, display a marked preference for lottery-type stocks (Frino et al., 2019; Lepone et al., 2023). Consequently, investment behaviours spurred by gambling mentalities within the Chinese stock market provide a unique perspective on the motivations underpinning investors' gambling proclivities. Considering China's progressive liberalisation of its capital market, elucidating the investment patterns and psychological undercurrents manifested in the Chinese market is imperative for foreign investors.

We select the Chinese market as our research sample and employ the Baidu search index of lottery-related keywords to measure the intensity of public gambling sentiment. We conduct an empirical analysis of how shifts in gambling sentiment influence the trading dynamics, liquidity and returns of lottery-type stocks, thereby discerning whether wealth acquisition or thrill seeking is the driving force behind investors' gambling mentality. To identify lottery-type stocks, we amalgamate the methodologies of Kumar (2009) and Bali et al. (2017) and utilise metrics such as idiosyncratic volatility, idiosyncratic skewness, nominal stock prices and historical peak returns. Furthermore, we categorise IPO stocks characterised by a significantly positive return distribution as a distinct subset of lottery-type stocks (Chen & Zheng, 2021). Our findings underscore the fact that Chinese investors perceive lotteries and lottery-type stocks as interchangeable investment avenues, with a preference for trading lottery-type stocks driven by wealth accumulation rather than the pursuit of excitement. Variations in public gambling sentiment intensity directly correlate with an uptick in the selling volume of lottery-type stocks and have an inverse relationship with short-term order imbalance, liquidity and the rate of abnormal returns. Additionally, lottery-type stocks with minimal institutional holdings are markedly more susceptible to fluctuations in gambling sentiment, suggesting the predominant influence of individual investors with gambling propensities as opposed to the lesser impact on institutional investors (Zhong, 2021). Among the spectrum of lottery-type stocks, Special Treatment (ST) stocks exhibit heightened sensitivity to gambling sentiment shifts.

This study makes three contributions to the theoretical literature. First, our empirical findings demonstrate that the predominant gambling mentality among Chinese investors is rooted in the pursuit of wealth. This offers compelling empirical evidence and pioneers a fresh theoretical perspective on what drives investors' gambling mentality, addressing a notable gap in the literature on the underlying causes of such a mentality. Second, leveraging the methods used by Chen et al. (2021), we apply fluctuations in Internet search indices to discern shifts in public gambling sentiment. From this vantage point, we probe the ramifications of these shifts in sentiment in terms of market trading patterns, liquidity and returns. Building upon the exploration of how gambling mentality influences overarching market trading behaviours (Dorn et al., 2015), our work can provide a theoretical explanation at the level of psychological motivation for market anomalies related to gambling sentiment such as the World Cup Curse, adding depth to the theoretical discourse on the repercussions of gambling inclinations. Finally, our study augments the body of work in behavioural finance pertaining to market pricing anomalies. While earlier research has alluded to pricing discrepancies such as the IPO pricing quirks highlighted by Chen and Zheng (2021), which are possibly linked to investors' gambling proclivities (Bergsma & Tayal, 2019), we emphasise that during heightened gambling sentiment, the pronounced IPO premium anomaly diminishes. This finding supports the assertion that a gambling mentality may be instrumental in IPO pricing irregularities, thereby presenting a novel empirical foundation for extant theoretical discourses.

2 THEORETICAL ANALYSIS AND HYPOTHESES DEVELOPMENT

2.1 Gambling mentality and financial decisions

Gambling mentality is often defined as the excessive pursuit of risk and misunderstanding or neglect of probability (Dorn et al., 2015). As investors are often faced with uncertainty and risk, this psychological phenomenon is particularly evident in financial decisions, usually manifested as overconfidence, risk preference, the pursuit of instant gratification and other characteristics (Ida & Goto, 2009). Investors with gambling mentality pay too much attention to short-term returns and ignore long-term risks and are more inclined to adopt high-risk investment strategies or even make irrational decisions, such as excessive trading, chasing the rise and killing the fall, and chasing hot spots in the market (Gong et al., 2021; Mosenhauer et al., 2021).

In existing research, scholars often attribute the individual motivation of investors' gambling mentality to cognitive biases, particularly misjudgements of probability caused by biases such as overconfidence and the illusion of control (Gong & Zhu, 2019). However, gambling mentality does not necessarily stem from cognitive bias; some gamblers still participate in gambling activities even if they can clearly recognise the risk and loss probability, which is seldom mentioned in financial decision-making research (Stetzka & Winter, 2021).

Research in related fields suggests that the pursuit of wealth and thrill seeking are underlying individual motivations that lead to gambling behaviour and mentality. On the one hand, influenced by materialistic cultures and personal economic pressures, some individuals hope to quickly acquire substantial wealth through gambling (Kumar et al., 2016). On the other hand, high-risk gambling activities are attractive to those seeking excitement and adventure because of their unpredictability and potential for high returns to serve as sources of stimulation (Gao & Lin, 2015). Therefore, this study analyses the impact of changes in gambling sentiment on lottery-type stock trading behaviour and returns, focusing on whether the gambling mentality of Chinese investors stems from the pursuit of wealth or thrill seeking.

2.2 Effect on trading behaviour

Concerning trading dynamics, alterations in gambling sentiment contingent on distinct investor objectives may have divergent influences on the trading volume of lottery-type stocks. Investors motivated by wealth acquisition primarily seek to amend their financial standing (Kumar et al., 2016; Liu et al., 2020) by perceiving lottery and lottery-type stocks as interchangeable investment avenues. We postulate that during periods of intense gambling sentiment, these investors would allocate more capital towards lottery purchases while concurrently liquidating their lottery-type stock positions. Consequently, a negative correlation emerges between shifts in gambling sentiment and both the acquisition volume and the order imbalance of lottery-type stocks, with a contrasting positive correlation observed in their disposal volume.

Conversely, thrill-seeking investors have an intricate trajectory. Participation in lottery and lottery-type stock markets represents an exhilarating venture. A potential substitution effect may surface, diminishing attention as individuals are lured from stock trades by the allure of lottery engagements, thereby curtailing their market transaction frequency (Brown et al., 2018; Huang et al., 2019). This dynamic suggests a negative correlation among lottery-type stock trading, buying and selling volumes, and oscillations in public gambling sentiment, with no marked influence on order imbalances. Alternatively, the spillover of lottery market gambling sentiment could permeate the stock market (Chen et al., 2021). Investors may prefer purchasing lottery tickets at the expense of other expenditures rather than curtailing alternate forms of gambling. In this context, we conjecture a positive relationship between the trading and acquisition volumes of lottery-type stocks and upticks in general gambling sentiment.

Given the above analysis, it is plausible that investors driven by varying gambling mentalities will demonstrate disparate trading patterns. Therefore, Hypotheses 1a, 1b and 1c are proposed.

H1a.For wealth-seeking investors, both the purchase volume and order imbalance of lottery-type stocks are negatively correlated with changes in gambling sentiment, whereas the selling volume shows a direct positive relationship.

H1b.Assuming a substitution effect, the trading, purchasing and selling volumes of lottery-type stocks are negatively correlated with changes in public gambling sentiment for thrill-seeking investors.

H1c.In a sentiment spillover scenario, both the trading and purchasing volumes of lottery-type stocks are positively correlated with changes in public gambling sentiment for thrill-seeking investors.

2.3 Effect on liquidity

Extant research underscores the fact that oscillations in gambling sentiment exert a tangible influence on the liquidity of lottery-type stocks (Meng & Pantzalis, 2018). Based on prior assumptions regarding trading behaviour, investors propelled by wealth-enhancement objectives may opt to liquidate lottery-type stock holdings, aiming for cash realisation at more advantageous prices, thereby augmenting the liquidity of the stock. This disposition could potentially be reflected as an inverse relationship between the bid-ask spread of lottery-type stocks and modifications in gambling sentiment. By contrast, thrill-seeking investors, influenced by escalating gambling sentiment, might curtail their stock market buy-and-sell activities, consequently impairing the liquidity of lottery-type stocks. This dynamic implies a direct positive correlation between the bid and ask spread of lottery-type stocks and fluctuations in public gambling sentiment. Nevertheless, variations in gambling sentiment can also prompt the unidirectional trading of lottery-type stocks, analogous to the strategies employed by wealth-motivated investors, enhancing liquidity and constricting the bid-ask spread of these stocks.

In light of these considerations, we propose Hypotheses 2a and 2b:

H2a.For wealth-seeking investors, or in instances of sentiment spillover effects among thrill-seeking participants, alterations in gambling sentiment correlate positively with liquidity and inversely with the bid-ask spread of lottery-type stocks.

H2b.For thrill-seeking investors experiencing a substitution effect, variations in gambling sentiment are negatively associated with liquidity, paralleled by a positive association with the bid-ask spread of lottery-type stocks.

2.4 Effect on returns

Existing research suggests that a prevalent gambling mentality influences the pricing of lottery-type stocks, often leading to short-term overvaluation, and ultimately, negative expected excess returns (Bali et al., 2017). This dynamic suggests that shifts in gambling sentiment can momentarily put pressure on prices, engendering abnormal returns on lottery-type stocks. In particular, for investors whose strategies are shaped by wealth-seeking motives, heightened gambling sentiment typically correlates with a reduction in lottery-type stock positions, precipitating negative short-term abnormal returns. Conversely, if trading behaviour has emotional spillovers, intensified gambling sentiment can amplify investors' risk tolerance, bolster the demand for lottery-type stocks, and foster positive short-term abnormal returns. These theoretical premises extend to specific categories of lottery-type stocks including IPOs. For instance, wealth-motivated investors captivated by the lottery market might divert their focus from IPOs, contributing to subdued first-day returns during periods of robust gambling sentiment. In contrast, thrill-seeking investors may exhibit greater enthusiasm for IPO acquisitions amid rising gambling sentiment, prompting elevated first-day returns (Chen & Zheng, 2021).

Hence, we propose Hypotheses 3a, 3b, 4a and 4b.

H3a.For wealth-seeking investors, a higher gambling sentiment results in lower short-term abnormal returns for lottery-type stocks.

H3b.For thrill-seeking investors, if a sentiment spillover effect exists, higher gambling sentiment will result in higher short-term abnormal returns for lottery-type stocks.

H4a.For wealth-seeking investors, a high gambling sentiment results in lower first-day returns on IPO stocks.

H4b.For thrill-seeking investors, if there is a sentiment spillover effect, high gambling sentiment will result in higher first-day returns on IPO stocks.

Figure 1 presents a logical framework for the change in gambling sentiment, which impacts the trade behaviour, liquidity and returns of lottery-type stocks.

Details are in the caption following the image — **FIGURE 1**
Open in figure viewer PowerPoint

Research framework.

3 METHODS AND DATA

3.1 Measuring public gambling sentiment

We apply the method of Chen et al. (2021) and use the Baidu index of lottery-related search terms as a barometer for public gambling sentiment. The Baidu index offers two significant advantages. First, it encompasses broader public engagement with lottery activities, presenting a more holistic view than studies confined to singular events such as lottery jackpots (Dorn et al., 2015). Second, Baidu-based sentiment indicators benefit from daily granularity, allowing for a more nuanced understanding of temporal shifts in gambling sentiment, an aspect not captured by other monthly indicators such as lottery sales. Certainly, there are limitations to using the Baidu index as a measure of gambling sentiment, as it is an indirect measure and not as direct as measuring investors' actual lottery purchasing behaviour data or local lottery sales data at the city level. However, because the latter two types of data are difficult to obtain in the Chinese market, this study uses the Baidu index to measure public gambling sentiment.

To construct a robust measure of public gambling sentiment, we identified five keywords central to lottery activities with substantial search volumes (‘lottery’, ‘welfare lottery’, ‘sports lottery’, ‘dual-coloured ball’ and ‘super-lotto’). The aggregation of the search indices for these terms constitutes the search volume intensity (SVI), our primary sentiment gauge. Subsequently, we derived the adjusted search volume index (ASVI) to serve as the pivotal independent variable, signifying real-time fluctuations in gambling sentiment intensity. In line with our research objectives, we opted for weekly data frequencies compiled through daily data aggregation.

{ASVI}_{t} = {LogSVI}_{t} - {LogSVI}_{t - 1}

()

In Equation (1),

{LogSVI}_{t}

and

{LogSVI}_{t - 1}

are the logarithms of the search volume intensity of gambling sentiment in periods t and t–1.

3.2 Identifying lottery-type stocks

We obtain the lottery characteristic index (LIDX) of each stock by arithmetically averaging the four indicators proposed by Kumar (2009) and Bali et al. (2017): idiosyncratic volatility, idiosyncratic skewness, nominal stock price and the historical highest return. The top 20% of

{LIDX}_{i}

stocks are defined as lottery-type stocks for the quarter, the bottom 20% as non-lottery-type stocks, and the remainder as other stocks.

{LIDX}_{i} = \frac{R_{vol, i} + R_{skew, i} + R_{price, i} + R_{MaxRet, i}}{4 \times N}

()

In Equation (2),

R_{vol, i}

is the descending order of stocks' idiosyncratic volatility,

R_{skew, i}

is the descending order of stocks' idiosyncratic skewness,

R_{price, i}

is the order of stocks' nominal prices,

R_{MaxRet, i}

is the highest historical return of stocks sorted in descending order, and N is the number of stock samples in the previous quarter.

3.3 Research design

We focus on the impact of changes in gambling sentiment on the trading behaviour of lottery-type stocks and construct a time-series regression model at the overall level of lottery-type stocks and a panel model at the individual testing level. The specific models are as follows.

{Trade}_{t} = β_{0} + β_{1} {ASVI}_{t - 1} + β_{2} {AvgTrade}_{[t - 1, t - 4]} + Controls + ε_{t}

()

{Trade}_{i, t} = β_{0} + β_{1} {ASVI}_{i, t - 1} + β_{2} {AvgTrade}_{i, [t - 1, t - 4]} + Controls + ε_{i, t}

()

We use the lottery-type stock trading volume (TrdNum), active buying volume (BNum), active selling volume (SNum) and order imbalance (BSI) as dependent variables (Trade) to describe the trading behaviour and direction of lottery-type stocks. We use ASVI, the change in the intensity of public gambling sentiment constructed above, as the core independent variable and process it with a lag of one period. We introduce the average value of the dependent variables with a lag of 4 weeks to a lag of 1 week (

{AvgTrade}_{[t - 1, t - 4]}

) to control for the time autocorrelation of the dependent variables. In addition, we control for factors that may affect lottery-type stock trading, such as market factors (market average return, market liquidity and market total circulation market value), economic factors (business index of macroeconomics, urban unemployment and interbank offered rate), and individual stock factors (stock return, stock liquidity and circulation market value), especially for panel models (Gao & Lin, 2015).

The measure of the imbalance in individual stocks' orders refers to the research of Bongaerts et al. (2022) and Zhang et al. (2020), and the arithmetic average of all lottery-type stocks is used as the overall level of the order imbalance index:

{BSI}_{i, t} = \frac{{BNum}_{i, t} - {SNum}_{i, t}}{{BNum}_{i, t} + {SNum}_{i, t}} .

()

In Equation (5),

{BNum}_{i, t}

and

{SNum}_{i, t}

are the active buying and selling volumes, respectively.

We focus on the impact of changes in gambling sentiment on the liquidity of lottery-type stocks and conduct an empirical test on the overall time series and panel data of individual stocks. We refer to the method of Hendershott et al. (2011) and Katselas et al. (2020), using the intraday time-weighted bid-ask spread to measure the liquidity level of the stock, including the relative quotation spread (Qsp) and the relative effective spread (Esp). We then use the time interval between two adjacent intraday transaction records as the weighting of the spread, and take the average of the calculation results for each trading day in week t to obtain Qsp_Time and Esp_Time as dependent variables that measure the overall or individual stock liquidity (Liq) of lottery-type stocks, respectively. The control variables are similar to those above except that the total market transaction amount (individual stocks) is used to replace liquidity. The model is

{Liq}_{t} = β_{0} + β_{1} {ASVI}_{t - 1} + β_{2} {AvgLiq}_{[t - 1, t - 4]} + Controls + ε_{t}

()

{Liq}_{i, t} = β_{0} + β_{1} {ASVI}_{i, t - 1} + β_{2} {AvgLiq}_{i, [t - 1, t - 4]} + Controls + ε_{i, t} .

()

In view of the influence of gambling sentiment on the return of lottery-type stocks, we first estimate the excess return of stock i in the window lagged by half a year:

r_{i, t} - r_{f, t} = α_{i} + β_{i} (r_{m, t} - r_{f, t}) + ε_{i, t .}

()

In Equation (8), $r_{i, t}$ is the return on stock i in week t, $r_{f, t}$ is the risk-free return in week t and $r_{m, t}$ is the market return in week t.

After obtaining the estimated results of

α_{i}

and

β_{i}

, as

{\hat{α}}_{i}

and

{\hat{β}}_{i}

, we calculate the abnormal return of stock i in week t as

{AR}_{i, t} = (r_{i, t} - r_{f, t}) - [{\hat{α}}_{i} + {\hat{β}}_{i} (r_{m, t} - r_{f, t})] .

()

Furthermore, we weigh the abnormal return of each lottery-type stock in week t using market capitalisation to obtain the abnormal return

{AR}_{t}

of the overall lottery-type stock in week t. We use

{AR}_{t}

and

{AR}_{i, t}

as dependent variables for the overall time-series and individual stock panel models.

{AR}_{t} = β_{0} + β_{1} {ASVI}_{t - n} + Controls + ε_{t}

()

{AR}_{i, t} = β_{0} + β_{1} {ASVI}_{i, t - n} + Controls + ε_{i, t}

()

The independent variable is

{ASVI}_{t - n}

, which represents the change in the intensity of public gambling sentiment with a lag of n weeks, n = 1, 2, 3, 4, and is used to explore how changes in gambling sentiment affect changes in stock prices during a period. The control variables are similar to those above except that the market (individual stocks) factors include liquidity, circulation market value and total trading volume.

Finally, we use the following regression model to test the impact of changes in gambling sentiment on the first-day returns on IPO stocks:

R_{i} = β_{0} + β_{1} {ASVI}_{i, t - 1} + Controls + ε_{i} .

()

In Equation (12), the dependent variable

R_{i}

is the return on IPO stock i on the first listing day, that is, the rate of change in the closing price relative to the first listing price. In addition to market factors, company age, issue price, price-to-earnings ratio and oversubscription multiples from the stock's IPO are controlled for.

3.4 Data and descriptive statistics

Appendix provides the definitions of the main variables. The sample period of this study is from 3 January 2011 to 28 March 2021 and excludes data after market close, for a total of 516 weeks and 254,589 lottery-type stock observations. The data sources used in this study are the Baidu, Wind and China Stock Market & Accounting Research (CSMAR) databases. Table 1 reports the basic characteristics of lottery-type- and non-lottery-type stocks. Compared with non-lottery-type stocks, lottery-type stocks have higher idiosyncratic volatility, idiosyncratic skewness, maximum returns and lower nominal prices, which is in line with the identification rules of lottery-type stocks and previous research. Table 2 reports the descriptive statistics of the core variables at the overall level for lottery-type stocks during the sample period.

TABLE 1. Comparison of basic characteristics of different types of stocks.

Variable	Lottery-type stocks	Non-lottery-type stocks	The between-group difference t-value
Trait volatility	5.340	2.392	39.789***
Trait skewness	1.544	−0.021	1.849*
Nominal share price	11.537	20.319	−33.789***
Highest return in history	7.245	3.560	99.031***
Institutional investor shareholding ratio	40.574	46.841	−26.398***
Logarithmic market capitalisation	15.256	15.481	−19.498***
Intraquarter earnings volatility	3.607	2.153	45.767***
Quarterly turnover rate of outstanding shares	207.006	97.535	77.636***
Book-to-market ratio	0.600	0.641	−16.475***

***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

TABLE 2. Descriptive statistics.

Variable	N	Mean	Std	Min	Median	Max
ASVI	516	0.006	0.200	−1.371	0.003	1.119
TrdNum	516	19.741	12.360	4.351	16.523	86.706
BNum	516	9.529	6.170	2.030	7.971	44.021
SNum	516	9.578	5.813	2.234	8.143	40.667
BSI	516	0.134	1.720	−4.048	0.028	6.822
Qsp_time	516	0.182	0.040	0.102	0.182	0.288
Esp_time	516	0.182	0.038	0.109	0.183	0.280
AR	516	−0.610	1.696	−8.808	−0.514	9.090
R	2298	0.594	0.789	−0.453	0.440	10.614

4 EMPIRICAL RESULTS

4.1 Baseline results

The results presented in Table 3 confirm Hypothesis 1a. We show that at the overall level, the total trading volume and active selling volume of lottery-type stocks are positively correlated with gambling sentiment, whereas active buying volume has no significant relationship with gambling sentiment. Meanwhile, the change in gambling sentiment was significantly negatively correlated with the order imbalance, which means that the proportion of active payments decreased, while the proportion of selling orders increased, resulting in an imbalance in buying and selling orders. At the individual stock level, active selling volume is positively correlated with gambling sentiment, and order imbalance is negatively correlated with gambling sentiment. The results indicate that investors with a gambling mentality mainly pursue wealth rather than thrills. Investors regard lottery and lottery-type stocks as alternative investments. During bullish periods, people choose to buy lottery-type stocks and actively sell and reduce their positions on them. This finding also confirms the views of Barberis and Huang (2008), who point out that gambling investors who pursue skewness deliberately maintain a single asset portfolio because diversification leads to a decrease in skewness. If people buy lottery-type stocks and trade them for thrill-seeking motives, they no longer need to seek stimulation from frequent buying and selling when they experience sufficient excitement in the lottery market (Dorn et al., 2015). Additionally, this phenomenon might be intertwined with the attention mechanism. Given that investors' capacity for attention is constrained, and in their perspective, both lottery-type stocks and lottery tickets are considered interchangeable avenues for wealth accumulation, they are inclined to diminish their stock holdings and redirect their capital and focus towards lottery trading endeavours.

TABLE 3. The impact of changes in gambling sentiment on lottery-type stock trading behaviour.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
Variable	TrdNum _t	BNum _t	SNum _t	BSI _t	SNum _i,t	BSI _i,t
ASVI _t–1	2.042*	0.817	1.032**	−0.691**	1.138***	−0.793***
ASVI _t–1	(1.85)	(1.51)	(1.98)	(−2.13)	(12.30)	(−10.00)
AvgTrdNum	0.876***
AvgTrdNum	(39.06)
AvgBNum		0.880***
AvgBNum		(40.57)
AvgSNum			0.878***		0.817***
AvgSNum			(38.46)		(570.60)
AvgBSI				0.687***		0.281***
AvgBSI				(12.54)		(83.80)
PRet	0.486***	0.334***	0.194***	0.114**	0.245***	0.086***
PRet	(3.16)	(4.43)	(2.66)	(2.52)	(82.60)	(34.03)
ILLIQ					0.626***	−0.573***
ILLIQ					(8.71)	(−9.30)
Wsmvosd					0.013***	0.001***
Wsmvosd					(59.37)	(8.15)
MktRet	−0.095	−0.103	−0.035	−0.061	−0.119***	−0.032***
MktRet	(−0.47)	(−1.04)	(−0.36)	(−1.03)	(−16.16)	(−5.15)
AILLIQ	0.566*	0.283**	0.241*	0.047	0.081***	0.023
AILLIQ	(1.94)	(1.97)	(1.74)	(0.54)	(3.49)	(1.16)
Cwmvosd	0.058**	0.026**	0.029**	−0.000	0.029***	0.008***
Cwmvosd	(2.25)	(2.13)	(2.36)	(−0.03)	(13.81)	(4.23)
Asset					−0.079***	−0.026***
Asset					(−8.53)	(−3.29)
Income					0.117***	0.047***
Income					(8.54)	(3.98)
Pnet					−4.577***	−0.859***
Pnet					(−23.71)	(−5.24)
Lev					−0.012	0.047***
Lev					(−0.85)	(3.83)
Shibor	0.362	0.184	0.184	−0.079	0.109***	−0.048**
Shibor	(1.26)	(1.31)	(1.36)	(−0.93)	(4.00)	(−2.04)
Macro	−0.240**	−0.116**	−0.109**	−0.016	−0.076***	−0.034***
Macro	(−2.23)	(−2.20)	(−2.13)	(−0.49)	(−8.97)	(−4.62)
UnEmpl	0.065	0.030	0.022	−0.023	0.047***	−0.049***
UnEmpl	(0.49)	(0.47)	(0.35)	(−0.59)	(4.31)	(−5.27)
Constant	−2.700*	−1.392*	−1.251	0.247	−0.970***	−0.246*
Constant	(−1.66)	(−1.74)	(−1.62)	(0.51)	(−6.35)	(−1.88)
Month fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Individual fixed effects	—	—	—	—	Yes	Yes
N	516	516	516	516	254,516	254,516
R ²	0.850	0.855	0.848	0.331	0.635	0.037

Note: For brevity, this table reports only the empirical results of the active selling volume (SNum) and order imbalance (BSI), which are closely related to our conclusions. Variable definitions are presented in the Appendix. T-statistics are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

The results presented in Table 4 confirm Hypothesis 2a. The coefficients of the changing intensity of gambling sentiment are significantly negative for both the overall and individual stock levels of lottery-type stocks, indicating that gambling sentiment has a significant and positive effect on the liquidity of lottery-type stocks. This again proves that the gambling mentality is driven by the pursuit of wealth. When individuals perceive lottery-type stocks and lotteries as being interchangeable in their quest for excitement, a shift in the gambling mood is unlikely to significantly influence market liquidity. However, when they regard these forms of speculation as substitutable on the level of accumulating wealth, an intense gambling sentiment, may lead investors to collectively offload stocks in favour of purchasing lotteries, fuelled by herd behaviour, consequently impacting market liquidity.

TABLE 4. The impact of changes in gambling sentiment on the liquidity of lottery-type stocks.

Variable	(1)	(2)	(3)	(4)
Variable	Qsp_time _t	Esp_time _t	Qsp_time _i,t	Esp_time _i,t
ASVI _t–1	−0.774**	−0.732**	−0.498***	−0.442***
ASVI _t–1	(−2.38)	(−2.28)	(−13.98)	(−12.24)
AvgQsp_time	0.981***		0.990***
AvgQsp_time	(56.01)		(994.79)
AvgEsp_time		0.977***		0.988***
AvgEsp_time		(53.14)		(956.76)
PRet	−0.441***	−0.454***	−0.061***	−0.047***
PRet	(−9.71)	(−10.07)	(−52.81)	(−39.92)
Wmount			−0.017***	−0.015***
Wmount			(−34.39)	(−30.43)
Wsmvosd			0.001***	0.001***
Wsmvosd			(9.63)	(7.75)
MktRet	0.281***	0.299***	−0.179***	−0.193***
MktRet	(4.71)	(5.06)	(−63.34)	(−67.48)
Amount	−0.000	−0.000	−0.000**	−0.000**
Amount	(−0.05)	(−0.05)	(−2.49)	(−2.47)
Cwmvosd	0.003	0.003	0.007***	0.007***
Cwmvosd	(0.51)	(0.54)	(8.17)	(8.42)
Asset			−0.001	−0.000
Asset			(−0.28)	(−0.06)
Income			0.013**	0.012**
Income			(2.44)	(2.30)
Pnet			−0.473***	−0.448***
Pnet			(−6.33)	(−5.90)
Lev			−0.020***	−0.018***
Lev			(−3.57)	(−3.10)
Shibor	−0.034	−0.030	−0.051***	−0.045***
Shibor	(−0.40)	(−0.36)	(−4.92)	(−4.26)
Macro	−0.036	−0.035	−0.047***	−0.045***
Macro	(−1.18)	(−1.18)	(−14.91)	(−14.14)
UnEmpl	0.008	0.007	−0.001	−0.002
UnEmpl	(0.21)	(0.17)	(−0.15)	(−0.40)
Constant	−0.034	−0.030	0.477***	0.478***
Constant	(−0.40)	(−0.36)	(7.78)	(7.66)
Month fixed effects	Yes	Yes	Yes	Yes
Individual fixed effects	—	—	Yes	Yes
N	516	516	254,516	254,516
R ²	0.874	0.862	0.812	0.799

Note: Variable definitions are presented in the Appendix. T-statistics are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

The results presented in Table 5 confirm Hypothesis 3a. First, changes in gambling sentiment with a lag period are significantly and negatively correlated with the abnormal returns of lottery-type stocks. Second, given a prolonged lag period, the negative impact caused by changes in gambling sentiment declines sharply, the resulting mispricing is corrected until the fourth lag period, and changes in gambling sentiment are significantly and positively correlated with the abnormal returns of lottery-type stocks. These results confirm that investors' gambling mentality stems from their pursuit of wealth. When general gambling sentiment increases, gambling investors take the initiative to sell lottery-type stocks in exchange for cash to buy lotteries. When sentiment declines, investors choose to continue buying or holding lottery-type stocks, causing abnormal short-term pricing. This discovery also offers an indirect theoretical explanation for the World Cup Curse in the Chinese market. During the World Cup, gambling sentiment is high, leading to capital outflow from the stock market by investors with a gambling mentality to chase wealth, and a temporary dip in returns and the market index.

TABLE 5. The impact of changes in gambling sentiment on the abnormal returns of lottery-type stocks.

Variable	(1)	(2)	(3)	(4)	(5)
Variable	AR _t	AR _t	AR _t	AR _t	AR _i,t
ASVI _t–1	−0.880**				−0.667***
ASVI _t–1	(−2.33)				(−9.31)
ASVI _t–2		−0.201
ASVI _t–2		(−0.53)
ASVI _t–3			−0.157
ASVI _t–3			(−0.40)
ASVI _t–4				0.795**
ASVI _t–4				(2.10)
ILLIQ					0.364***
ILLIQ					(6.54)
Wmount					−0.014***
Wmount					(−14.10)
Wsmvosd					−0.000**
Wsmvosd					(−1.99)
AILLIQ	−0.067	−0.074	−0.070	−0.074	−0.018
AILLIQ	(−0.67)	(−0.74)	(−0.69)	(−0.74)	(−0.98)
Amount	0.000	0.000	0.000	0.000	0.002***
Amount	(0.18)	(0.17)	(0.16)	(0.15)	(6.18)
Cwmvosd	−0.001	−0.001	0.000	0.000	0.018***
Cwmvosd	(−0.18)	(−0.12)	(0.00)	(0.02)	(11.04)
Asset					0.013*
Asset					(1.81)
Income					−0.025**
Income					(−2.39)
Pnet					0.121
Pnet					(0.80)
Lev					0.028**
Lev					(2.52)
Shibor	−0.162	−0.153	−0.140	−0.145	0.068***
Shibor	(−1.65)	(−1.55)	(−1.41)	(−1.47)	(3.24)
Macro	0.016	0.016	0.014	0.008	0.003
Macro	(0.44)	(0.42)	(0.39)	(0.20)	(0.52)
UnEmpl	−0.012	−0.013	−0.014	−0.013	−0.026***
UnEmpl	(−0.27)	(−0.28)	(−0.30)	(−0.29)	(−3.03)
Constant	−0.162	−0.153	−0.140	−0.145	−1.341***
Constant	(−1.65)	(−1.55)	(−1.41)	(−1.47)	(−11.37)
Month fixed effects	Yes	Yes	Yes	Yes	Yes
Individual fixed effects	—	—	—	—	Yes
N	516	515	514	513	254,516
R ²	0.052	0.041	0.041	0.049	0.005

Note: Variable definitions are presented in the Appendix. T-statistics are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

For IPO stocks, which are special lottery-type stocks, we selected 2698 IPO companies during the sample period and further tested the impact of changes in gambling sentiment on first-day returns. The results presented in Table 6 confirm Hypothesis 4a. We find a significant negative correlation between IPO first-day returns and changes in gambling sentiment, which indicates that when gambling sentiment is strong, gambling investors pay attention to funds in the lottery market, thereby reducing IPO participation and resulting in low IPO first-day returns. Simultaneously, the findings provide new evidence on the relationship between investors' gambling mentality, skewness preference and IPO anomalies. IPO anomalies refer to the fact that stocks tend to have high gains and turnover rates on the first listing day and their long-term performance is generally sluggish (Guo et al., 2023). Chen and Zheng (2021) point out that this is because investors with a gambling mentality are willing to pay an additional premium for positively skewed stocks, and view IPOs as an opportunity for potentially high wealth returns, leading to the overpricing of IPOs on the first day. The findings of this study provide side evidence of investors' limited attention in behavioural finance. When public gambling sentiment is high, gambling investors' attention shifts from the stock market to the lottery market, thereby reducing the abnormal premium for first-day IPO returns.

TABLE 6. The impact of changes in gambling sentiment on first-day IPO returns.

Variable	(1)	(2)	(3)	(4)
Variable	R _i	R _i	R _i	R _i
ASVI _t–1	−0.200**	−0.250***	−0.209**	−0.210**
ASVI _t–1	(−2.19)	(−2.88)	(−2.50)	(−2.43)
Age		0.081**	−0.003	−0.002
Age		(2.36)	(−0.09)	(−0.06)
Prc		0.001	−0.015	−0.035
Prc		(0.04)	(−0.68)	(−1.59)
PE		0.002***	0.002***	0.001***
PE		(5.43)	(4.57)	(4.14)
OSMul		0.101***	0.027***	0.053***
OSMul		(11.65)	(2.66)	(4.74)
AILLIQ			−0.032**	−0.021
AILLIQ			(−2.22)	(−1.35)
Amount			−0.000	−0.000
Amount			(−1.07)	(−0.55)
Cwmvosd			0.018***	0.013***
Cwmvosd			(13.78)	(8.09)
Shibor				0.054**
Shibor				(2.51)
Macro				−0.003
Macro				(−0.47)
UnEmpl				0.049***
UnEmpl				(5.71)
Constant	0.593***	−0.377***	−0.257**	−0.441***
Constant	(36.05)	(−3.26)	(−2.31)	(−2.74)
Month fixed effects	—	—	—	Yes
N	2298	2274	2274	2274
R ²	0.002	0.080	0.152	0.181

Note: Variable definitions are presented in the Appendix. T-statistics are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

4.2 Heterogeneity analysis

Prior studies demonstrate that in contrast to institutional investors, individual investors (Frino et al., 2019; Kumar, 2009; Lepone et al., 2023) prefer stocks with lottery features and that a gambling mentality makes individual investors tend to hold many stocks with high skewness. Institutional investors mostly exhibit a rational and stable investment philosophy and make few gambling-related investments. Therefore, we speculate that the impact of changes in gambling sentiment on lottery-type stocks is driven primarily by individual investors. Therefore, lottery-type stocks with a higher proportion of individual investors are more strongly affected than other stocks.

Furthermore, we introduce a grouping dummy variable 𝑖𝑛𝑠𝑡_i in the regression at the individual stock level, which takes the value of 1 when the shareholding ratio of institutional investors in lottery-type stock i ranked in the top 25% at the end of the last quarter and 0 when ranked in the last 25%. We tested the multiplicative coefficients of changes in gambling sentiment for heterogeneity. Table 7 reports the empirical results. All the multiplier coefficients are significant, indicating that a higher institutional shareholding ratio weakens the influence of the changes in gambling sentiment on lottery-type stocks and the abnormal return. This result fully shows that the impact of changes in public gambling sentiment on lottery-type stocks is dominated mainly by individual investors in the Chinese market, as they are more likely to have a gambling mentality and are more easily affected by changes in gambling sentiment.

TABLE 7. Heterogeneity test of institutional investors' shareholding ratio.

Variable	(1)	(2)	(3)	(4)
Variable	BSI _i,t	Qsp_time _i,t	Esp_time _i,t	AR _i,t
ASVI _t–1	−1.368***	−0.758***	−0.695***	−0.483***
ASVI _t–1	(−8.44)	(−11.11)	(−10.02)	(−3.47)
Inst × ASVI _t–1	1.556***	0.513***	0.477***	−0.424**
Inst × ASVI _t–1	(6.84)	(5.36)	(4.89)	(−2.17)
Controls	Yes	Yes	Yes	Yes
Constant	Yes	Yes	Yes	Yes
Month fixed effects	Yes	Yes	Yes	Yes
Individual fixed effects	Yes	Yes	Yes	Yes
N	117,090	117,090	117,090	117,090
R ²	0.037	0.812	0.799	0.005

Note: For brevity, this table reports only the regression results of the core variables. Variable definitions are presented in the Appendix. T-statistics are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

Existing studies have also noted that a gambling mentality can lead to the abnormal pricing of stocks of companies in financial distress or those that have uncertain prospects (Eraker & Ready, 2015). In China, ST companies are often in trouble, and therefore have higher risks. However, at the same time, there is also the expectation of decapping or reorganisation, and a few ST companies experienced a substantial increase in their stock prices after turning losses into profits. Thus, ST stocks exhibit obvious lottery characteristics that fit people's gambling attitudes and preferences. We further explore whether ST stocks moderate the influence of changes in gambling sentiment. We speculate that, among lottery-type stocks, ST stocks are more strongly affected by changes in gambling sentiment. Specifically, we introduce the grouping dummy variable ST in the regression at the individual stock level, which takes the value of 1 if lottery-type stock i is a special treatment stock at the end of the last week, including ST and *ST, and 0 otherwise. We tested the multiplicative coefficients of changes in gambling sentiment for heterogeneity. Considering the comparability of the samples, we excluded those that always had ST or non-ST status within the interval. Table 8 reports the empirical results and shows that the crossover item is significantly negatively correlated with the order imbalance and the bid-ask spread, indicating that when gambling sentiment is higher, the order imbalance and liquidity of ST stocks increase more than otherwise, reflecting the stronger lottery familiarity with ST stocks. In addition, we see no significant difference in the abnormal returns of ST and non-ST lottery-type stocks due to changes in gambling sentiment, which may be caused by restrictions in the special trading rules that apply to increases and decreases in ST stock prices.

TABLE 8. Heterogeneity test for special treatment (ST) stocks.

Variable	(1)	(2)	(3)	(4)
Variable	BSI _i,t	Qsp_time _i,t	Esp_time _i,t	AR _i,t
ASVI _t–1	−0.992***	−0.601***	−0.558***	−0.376*
ASVI _t–1	(−4.98)	(−4.61)	(−4.24)	(−1.85)
ST × ASVI _t–1	−0.779*	−0.588**	−0.586**	−0.255
ST × ASVI _t–1	(−1.90)	(−2.20)	(−2.17)	(−0.61)
Controls	Yes	Yes	Yes	Yes
Constant	Yes	Yes	Yes	Yes
Month fixed effects	Yes	Yes	Yes	Yes
Individual fixed effects	Yes	Yes	Yes	Yes
N	50,322	50,322	50,322	50,322
R ²	0.074	0.844	0.836	0.011

Note: For brevity, this table reports only the regression results of the core variables. Variable definitions are presented in the Appendix. T-statistics are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

4.3 Robustness tests

We conduct four endogeneity and robustness tests. First, we replace some of the dependent variables to avoid measurement errors. To measure lottery-type stock liquidity, the spread index weighted by the order interval time is replaced with the equal-weight spread index (Qsp_equal and Esp_equal) and the spread index weighted by the transaction amount (Qsp_amt) and (Esp_amt). To measure the abnormal returns of lottery-type stocks, the half-year rolling window period for calculating abnormal returns was replaced with a quarter-year rolling window (AR_{[t–13, t–1]}). The first-day return of the IPO stock is adjusted by the market return as in ${adR}_{i} = (1 + R_{i}) / (1 + R_{m}) - 1$ , where R_m is the market return on that day. As Table 9 shows, the empirical results after replacing the dependent variables are consistent with those of the main regression and are robust.

TABLE 9. Robustness test: Replacing the dependent variables.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
Variable	Qsp_equal _t	Esp_equal _t	Qsp_amt _t	Esp_amt _t	AR _{[t–13, t–1]}	adR _i
ASVI _t–1	−0.747**	−0.727**	−0.821**	−0.779**	−0.795**	−0.204**
ASVI _t–1	(−2.57)	(−2.45)	(−2.26)	(−2.00)	(−2.07)	(−2.37)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Constant	Yes	Yes	Yes	Yes	Yes	Yes
Month fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
N	516	516	516	516	516	2274
R ²	0.891	0.878	0.850	0.816	0.045	0.179

Note: For brevity, this table reports only the regression results of the core variables. Variable definitions are presented in the Appendix. T-statistics are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

Second, considering the possible endogenous interference of market investor sentiment in the empirical results, we introduce the investor sentiment index of the Chinese stock market (ISI) as an additional control variable to address the endogenous influence of overall market sentiment. Third, the closing of the lottery market during the Spring Festival and National Day may cause changes in gambling sentiment. The Spring Festival holiday also has a holiday effect on stock market transactions (Doran et al., 2012), possibly creating endogeneity problems. Although the potential cyclical effect has been mitigated by controlling for the month dummy variable in the main regression, to further address the related endogeneity problem, we exclude special samples during the Spring Festival and National Day from the robustness test. As Table 10 indicates, the empirical results remain robust.

TABLE 10. Robustness test: Adding investor sentiment control variables or excluding special samples during festivals.

Variable	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Variable	BSI _t	Qsp_time _t	Esp_time _t	AR _t	BSI _t	Qsp_time _t	Esp_time _t	AR _t
ASVI _t–1	−0.669**	−0.008**	−0.007**	−0.866**	−0.868**	−0.008*	−0.008*	−1.400***
ASVI _t–1	(−2.09)	(−2.42)	(−2.31)	(−2.32)	(−2.12)	(−1.93)	(−1.90)	(−2.97)
ISI _t–1	0.336***	0.003***	0.003***	−0.329***
ISI _t–1	(3.35)	(3.35)	(3.32)	(−3.20)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Month fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	516	516	516	516	491	491	491	491
R ²	0.344	0.876	0.865	0.065	0.348	0.878	0.866	0.063

Note: For brevity, this table reports only the regression results of the core variables. Variable definitions are presented in the Appendix. T-statistics are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

Fourth, we address the endogeneity problem caused by possible common factors affecting both stock trading and gaming sentiment by comparing the differences in the impact of changes in gaming sentiment on lottery- and non-lottery-type stocks. We introduce the stock-type grouping dummy variable D, which takes the value of 1 when the stock is classified as a lottery-type stock, and 0 otherwise. We further explore the difference in the impact of changes in gambling sentiment on the two types of stocks by observing the multiplication term, which is constructed using the dummy variable D and the core independent variable ASVI. The results in Table 11 confirm the robustness of the main results. Except for the imbalance in individual stock orders, the coefficients on the multipliers are all significant, indicating that changes in gambling sentiment have a significantly greater impact on lottery-type stocks than on non-lottery-type stocks.

TABLE 11. Heterogeneous impact of changes in gambling sentiment on lottery- and non-lottery-type stocks.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
Variable	TrdNum _i,t	SNum _i,t	BSI _i,t	Qsp_time _i,t	Esp_time _i,t	AR _i,t
ASVI _t–1	0.780***	0.448***	−0.880***	−0.406***	−0.365***	−0.104
ASVI _t–1	(3.56)	(4.44)	(−8.22)	(−9.78)	(−8.83)	(−1.37)
D × ASVI _t–1	1.639***	0.729***	0.109	−0.204***	−0.183***	−0.505***
D × ASVI _t–1	(5.50)	(5.30)	(0.75)	(−3.60)	(−3.26)	(−4.91)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Constant	Yes	Yes	Yes	Yes	Yes	Yes
Month fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Individual fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
N	378,965	378,965	378,965	378,965	378,965	378,965
R ²	0.646	0.648	0.045	0.778	0.766	0.007

Note: For brevity, this table reports only the regression results of the core variables. Variable definitions are presented in the Appendix. T-statistics are reported in parentheses. ***, ** and * indicate statistical significance at the 1%, 5% and 10% levels, respectively.

5 CONCLUSION AND IMPLICATIONS

Extant research underscores the prevalence of a gambling mentality in equity markets, manifesting as inflated short-term returns and diminished long-term yields for lottery-type stocks (Barberis & Huang, 2008). Recognising the temporal dynamics of public gambling sentiment, this study leverages the Baidu index, which encompasses lottery-related keywords, to gauge shifts in this sentiment. Our empirical scrutiny of the influence of these shifts on lottery-type stocks aims to dissect the underpinnings of investors' speculative inclinations.

Wealth accumulation, rather than thrill seeking, fuels investors' gambling mentality, with lotteries and analogous stocks viewed as interchangeable investment avenues. The repercussions of fluctuating gambling sentiment are threefold: it catalyses a surge in the aggregate trade volume of lottery-type stocks, particularly in active sell-offs, leading to order imbalances; enhances the liquidity of such stocks when sentiment peaks; and constricts their abnormal returns, a constraint that gradually attenuates and inverts. Notably, escalating gaming sentiment depresses the inaugural returns of IPO stocks, a subset of lottery-type stocks. Moreover, we observe that individual investors, especially those engaged in ST stocks that display pronounced lottery features, are more susceptible to sentiment shifts than their institutional counterparts.

This study has significant implications for both market participants and regulators. It is imperative that investors enhance their self-awareness of their inherent biases towards risk-taking and understand how the pursuit of wealth influences their choices. They must prioritise the development of a strategic long-term investment plan aimed at effectively diversifying risk to achieve reliable returns, eschewing the allure of transient profits. Financial institutions are key players in nurturing an environment conducive to informed, considered investment approaches rather than fostering unrealistic expectations of swift, substantial gains. Concurrently, these institutions should ensure that their sales teams are adept at identifying and managing clients' gambling tendencies, thus preventing investor misguidance through responsible practices. Regulatory bodies must uphold their roles with vigilant oversight and robust consumer education, creating a fair and stable financial landscape in which speculative impulses are curtailed and investor interests are protected. Through concerted efforts, it is possible to guide investments away from a gambling mentality and towards sustainable and judicious financial behaviour.

Nevertheless, this study acknowledges the constraints rooted primarily in data accessibility, necessitating a reliance on aggregate market trends and yielding less precise insights. Subsequent research, facilitated by expanded data repositories, could offer granular analyses of individual trading accounts and lottery engagements, thus enhancing our understanding of investor psychology. Furthermore, considering the diverse lottery ecosystems in China, future studies could investigate how varying gambling subsegments influence market participation and dynamics. In addition, regional disparities in speculative tendencies, as Ji et al. (2021) document, suggest a potential avenue for exploring geographical nuances in sentiment shifts.

ACKNOWLEDGEMENTS

This work is supported by the National Natural Science Foundation of China (No. 72001104, 72371125, 72172020) and the Social Science Foundation of Jiangsu Province (No. 22EYC002).

APPENDIX: VARIABLE DEFINITIONS

Variables	Symbol	Definitions
Dependent variables
Total transaction volume	TrdNum	Total turnover of individual stocks or the arithmetic average of lottery-type stocks
Active buying volume	BNum	Number of active purchases of individual stocks or the arithmetic average of lottery-type stocks
Active buying volume	SNum	Number of actively sold shares of individual stocks or the arithmetic average of lottery-type stocks
Order imbalance	BSI	Order imbalance index of individual stocks or the arithmetic mean of lottery-type stocks
Relative quote spread	Qsp_Time	Time-weighted relative spreads of individual stocks or arithmetic mean of lottery-type stocks
Relative effective spread	Esp_Time	Time-weighted absolute spread of individual stocks or arithmetic mean of lottery-type stocks
Abnormal return	AR	Abnormal return of individual stocks or the weighted average of the float-based market capitalisation of lottery-type stocks
First-day return of IPO stock	R	Rate of change in the closing price relative to the first listing price
Independent variables
Intensity of changes in public gambling sentiment	ASVI	Baidu search index change intensity of lottery-related keywords at week level
Control variables
Market average return	MktRet	Float-to-market-weighted average return of all stocks in the market
Market liquidity	AILLIQ	Amihud's liquidity indicators of all stocks in the market (Amihud, 2002)
Market transaction amount	Amount	The total transaction amount of all trading days in a week of all stocks in the market
Market total circulation market value	Cwmvosd	Total circulating market value of all stocks in the market within the week
Lottery-type stock return	PRet	Market capitalisation weighted average return of all lottery-type stocks or individual stock
Total asset size	Asset	Logarithm of the total asset size of an individual company
Total operating income	Income	Logarithm of the total operating income of an individual company
Net profit	Pnet	Logarithm of the net profit of an individual company
Asset liability ratio	Lev	Asset liability ratio of an individual company
Stock liquidity	ILLIQ	Amihud's liquidity indicators of individual stocks (Agarwal et al., 2022)
Stock transaction amount	Wmount	The total transaction amount of all trading days in a week of individual stocks
Stock circulation market value	Wsmvosd	Total circulating market value of individual stocks within the week
Business index of macroeconomic	Macro	Rate of change in the macroeconomic business index, which reflects the basic trend in the current economy and is composed of four aspects from the regular questionnaire survey of entrepreneurs conducted by the National Bureau of Statistics, including industrial production, employment, social demand and social income
Urban unemployment	UnEmpl	Rate of change in the urban unemployment rate
Interbank offered rate	Shibor	National Interbank 7-Day interbank offered rate
Age of the company	Age	Logarithm of the number of years a company was established at its IPO
Issue price	Prc	Logarithm of IPO price
Price–earnings ratio	PE	IPO price per share divided by earnings per share
Oversubscription multiples	OSMul	Logarithm of the number of shares subscribed by investors divided by the total number of shares issued at IPO

Open Research

DATA AVAILABILITY STATEMENT

Data will be made available on request.

REFERENCES

Agarwal, V., Jiang, L. & Wen, Q. (2022) Why do mutual funds hold lottery stocks? Journal of Financial and Quantitative Analysis, 57(3), 825–856.
10.1017/S0022109021000211
Web of Science® Google Scholar
Amihud, Y. (2002) Illiquidity and stock returns: cross-section and time-series effects. Journal of Financial Markets, 5(1), 31–56.
10.1016/S1386-4181(01)00024-6
Web of Science® Google Scholar
An, L., Wang, H.J., Wang, J. & Yu, J. (2020) Lottery-related anomalies: the role of reference-dependent preferences. Management Science, 66(1), 473–501.
10.1287/mnsc.2018.3205
Web of Science® Google Scholar
Bali, T.G., Brown, S.J., Murray, S. & Tang, Y. (2017) A lottery-demand-based explanation of the beta anomaly. Journal of Financial and Quantitative Analysis, 52(6), 2369–2397.
10.1017/S0022109017000928
Web of Science® Google Scholar
Barberis, N. & Huang, M. (2008) Stocks as lotteries: the implications of probability weighting for security prices. American Economic Review, 98(5), 2066–2100.
10.1257/aer.98.5.2066
Web of Science® Google Scholar
Bergsma, K. & Tayal, J. (2019) Short interest and lottery stocks. Financial Management, 48(1), 187–227.
10.1111/fima.12220
Web of Science® Google Scholar
Bongaerts, D., Roll, R., Rösch, D., van Dijk, M. & Yuferova, D. (2022) How do shocks arise and spread across stock markets? A microstructure perspective. Management Science, 68(4), 3071–3089.
10.1287/mnsc.2021.3979
Web of Science® Google Scholar
Brown, S., Lu, Y., Ray, S. & Teo, M. (2018) Sensation seeking and hedge funds. The Journal of Finance, 73(6), 2871–2914.
10.1111/jofi.12723
Web of Science® Google Scholar
Chen, H.H. & Zheng, M.R. (2021) IPO underperformance and the idiosyncratic risk puzzle. Journal of Banking & Finance, 131, 106190.
10.1016/j.jbankfin.2021.106190
Web of Science® Google Scholar
Chen, Y., Kumar, A. & Zhang, C.D. (2021) Searching for gambles: gambling sentiment and stock market outcomes. Journal of Financial and Quantitative Analysis, 56(6), 2010–2038.
10.1017/S0022109020000496
Web of Science® Google Scholar
Cheon, Y.H. & Lee, K.H. (2018) Making out globally: individualism, investor attention, and the cross section of expected stock returns. Management Science, 64(12), 5807–5831.
10.1287/mnsc.2017.2830
Web of Science® Google Scholar
Doran, J.S., Jiang, D.L. & Peterson, D.R. (2012) Gambling preference and the new year effect of assets with lottery features. Review of Finance, 16(3), 685–731.
10.1093/rof/rfr006
Web of Science® Google Scholar
Dorn, A.J., Dorn, D. & Sengmueller, P. (2015) Trading as gambling. Management Science, 61(10), 2376–2393.
10.1287/mnsc.2014.1979
Web of Science® Google Scholar
Drummond, P.A. (2023) Market quality surrounding anticipated distraction events: evidence from the FIFA World Cup. Journal of Financial Markets, 63, 100768.
10.1016/j.finmar.2022.100768
Web of Science® Google Scholar
Eraker, B. & Ready, M. (2015) Do investors overpay for stocks with lottery-like payoffs? An examination of the returns of OTC stocks. Journal of Financial Economics, 115(3), 486–504.
10.1016/j.jfineco.2014.11.002
Web of Science® Google Scholar
Fjesme, S.L., Lv, J.R. & Shekhar, C. (2023) The world cup in football and the US IPO market. Journal of Corporate Finance, 80, 102410.
10.1016/j.jcorpfin.2023.102410
Web of Science® Google Scholar
Frino, A., Lepone, G. & Wright, D. (2019) Are paper winners gamblers? Evidence from Australian retail investors. Accounting & and Finance, 59, 593–614.
10.1111/acfi.12296
Web of Science® Google Scholar
Gao, X.H. & Lin, T.C. (2015) Do individual investors treat trading as a fun and exciting gambling activity? Evidence from repeated natural experiments. The Review of Financial Studies, 28(7), 2128–2166.
10.1093/rfs/hhu075
Web of Science® Google Scholar
Gong, P., Wen, Z., Xiong, X. & Gong, C.M. (2021) When do investors gamble in the stock market? International Review of Financial Analysis, 74, 101712.
10.1016/j.irfa.2021.101712
Web of Science® Google Scholar
Gong, X. & Zhu, R. (2019) Cognitive abilities, non-cognitive skills, and gambling behaviors. Journal of Economic Behavior & Organization, 165, 51–69.
10.1016/j.jebo.2019.06.016
Web of Science® Google Scholar
Green, T.C. & Hwang, B.H. (2012) Initial public offerings as lotteries: skewness preference and first-day returns. Management Science, 58(2), 432–444.
10.1287/mnsc.1110.1431
Web of Science® Google Scholar
Guo, D.Z., Zheng, Y.Y., Wang, W.S., Hu, P.-N., Yang, Z. & Chen, Z. (2023) The impact of different sentiment in investment decisions: evidence from China's stock markets IPOs. Economic Research-Ekonomska Istrazivanja, 36(1), 2113739.
10.1080/1331677X.2022.2113739
Google Scholar
Hendershott, T., Jones, C.M. & Menkveld, A.J. (2011) Does algorithmic trading improve liquidity? The Journal of Finance, 66(1), 1–33.
10.1111/j.1540-6261.2010.01624.x
Web of Science® Google Scholar
Huang, S.Y., Huang, Y.L. & Lin, S.C. (2019) Attention allocation and return co-movement: evidence from repeated natural experiments. Journal of Financial Economics, 132(2), 369–383.
10.1016/j.jfineco.2018.10.006
Web of Science® Google Scholar
Ida, T. & Goto, R. (2009) Interdependency among addictive behaviours and time/risk preferences: discrete choice model analysis of smoking, drinking, and gambling. Journal of Economic Psychology, 30(4), 608–621.
10.1016/j.joep.2009.05.003
Web of Science® Google Scholar
Ji, Q., Quan, X.F., Yin, H.Y. & Yu, Q. (2021) Gambling preferences and stock price crash risk: evidence from China. Journal of Banking & Finance, 128, 106158.
10.1016/j.jbankfin.2021.106158
Web of Science® Google Scholar
Katselas, D., Sidhu, B.K. & Yu, C. (2020) Liquidity and information asymmetry around unscheduled mining announcements. Accounting & Finance, 61(2), 3053–3087.
10.1111/acfi.12694
Web of Science® Google Scholar
Kumar, A. (2009) Who gambles in the stock market? The Journal of Finance, 64(4), 1889–1933.
10.1111/j.1540-6261.2009.01483.x
Web of Science® Google Scholar
Kumar, A., Page, J.K. & Spalt, O.G. (2016) Gambling and comovement. Journal of Financial and Quantitative Analysis, 51(1), 85–111.
10.1017/S0022109016000089
Web of Science® Google Scholar
Lepone, G., Westerholm, J. & Wright, D. (2023) Speculative trading preferences of retail investor birth cohorts. Accounting & Finance, 63(1), 555–574.
10.1111/acfi.12934
Web of Science® Google Scholar
Li, X.D., Subrahmanyam, A. & Yang, X.W. (2018) Can financial innovation succeed by catering to behavioral preferences? Evidence from a callable options market. Journal of Financial Economics, 128(1), 38–65.
10.1016/j.jfineco.2018.01.010
Web of Science® Google Scholar
Liu, B.B., Wang, H.J., Yu, J.F. & Zhao, S. (2020) Time-varying demand for lottery: speculation ahead of earnings announcements. Journal of Financial Economics, 138(3), 789–817.
10.1016/j.jfineco.2020.06.016
Web of Science® Google Scholar
Meng, Y. & Pantzalis, C. (2018) Monthly cyclicality in retail investors' liquidity and lottery-type stocks at the turn of the month. Journal of Banking & Finance, 88, 176–191.
10.1016/j.jbankfin.2017.11.012
Web of Science® Google Scholar
Mosenhauer, M., Newall, P.W.S. & Walasek, L. (2021) The stock market as a casino: associations between stock market trading frequency and problem gambling. Journal of Behavioral Addictions, 10(3), 683–689.
10.1556/2006.2021.00058
PubMed Web of Science® Google Scholar
Perez, L. & Humphreys, B. (2013) The ‘who and why’ of lottery: empirical highlights from the seminal economic literature. Journal of Economic Surveys, 27(5), 915–940.
10.1111/j.1467-6419.2012.00732.x
Web of Science® Google Scholar
Stetzka, R.M. & Winter, S. (2021) How rational is gambling? Journal of Economic Surveys, 37(4), 1432–1488.
10.1111/joes.12473
Web of Science® Google Scholar
Tse, S., Yu, A.C.H., Rossen, F. & Wang, C.W. (2010) Examination of Chinese gambling problems through a socio-historical-cultural perspective. The Scientific World Journal, 10, 1694–1704.
10.1100/tsw.2010.167
PubMed Google Scholar
Zhang, T., Jiang, G.J. & Zhou, W.X. (2020) Order imbalance and stock returns: new evidence from the Chinese stock market. Accounting & Finance, 61(2), 2809–2836.
10.1111/acfi.12684
Web of Science® Google Scholar
Zhong, A. (2021) Institutional trading in stock market anomalies in Australia. Accounting & Finance, 62(1), 893–930.
10.1111/acfi.12813
Web of Science® Google Scholar
Zhu, H.B., Yang, L.H. & Xu, C.X. (2023) Tracking investor gambling intensity. International Review of Financial Analysis, 86, 102468.
10.1016/j.irfa.2022.102468
Web of Science® Google Scholar

Volume64, Issue5

December 2024

Pages 4515-4536

The individual motivation of investors' gambling mentality: Empirical evidence from gambling sentiment and lottery-type stocks

Abstract

1 INTRODUCTION

2 THEORETICAL ANALYSIS AND HYPOTHESES DEVELOPMENT