Borden(1942) put forward the dominant argument of advertising on behalf of the economics circle through the research on advertising promotion effect from 1919 to 1940, that is, the commodities and market information spread by advertising are conducive to the correct consumption of consumers and the sales of enterprise products. Subsequently, many economists took part in it one after another. From the perspective of consumers’ rights and interests, utility, and rational and optimal allocation of social and economic resources, they discussed and studied the welfare economy nature of advertising in large quantities.

Borden’s conclusion was based on the assumption that the total demand would increase, but with the gradual saturation of the market, the total demand for goods would maintain at a certain level. Kaldor (1945) first questioned the traditional view on the role of advertising and promotion “Proposed by Marshall as early as the beginning of the 20th century AD, especially persuasive advertisement is just the people to the demand for a product from one brand to another brand” point of view (that is, the brand advertising demand transfer function) is correct, and by establishing the model, the method of the AD is not yet fully developed in the economic life period is excessive point of view, and boldly put forward the number of ads for the welfare of the advertising economics thesis sets a preliminary economic analysis. Palda (1965) tested a series of advertising consumption models by multiple regression, and the results showed that advertising is an amortized asset.

On this basis, Palda proposed that advertising has a cumulative effect, providing a unique analytical perspective for the study of the relationship between advertising and sales. Economist Galbraith (1967) believed that advertising stimulates total demand and adapts to economic development by changing consumers’ behaviors. Since the development of economic system requires a long period, there is a long-term equilibrium relationship between advertising expenditure and overall sales (consumption). In accordance with such logic,Galbraith proposed the hypothesis that advertising stimulates sales. In the following research, the researcher used the data of enterprise level, industry level or national level to establish the model of advertising and sales (or consumption) for empirical test.

In terms of the correlation between advertising and sales, Aaker et al. (1982) conducted an empirical test on the monthly data of food brands and found that there was no significant relationship between advertising expenditure and sales. Chowdhury(1994) also found no stable relationship between advertising expenditure and sales through co-integration analysis and causal test of British data. Duffy (1996) used data from multiple sources to examine the relationship between advertising and sales from the industrial level and found that tobacco advertising did not stimulate the overall sales of tobacco. Based on this, Duffy believed that food advertising did not change the household consumption budget to stimulate the overall sales of food. Dekimpe et al. (1995) argued that advertising is only one of many tactical means to influence sales, and other factors have a greater stimulating effect on sales than advertising.

The co-integration relationship between advertising and sales only appears in specific circumstances, and it is difficult for them to have a long-term co-integration relationship due to the more unstable data at the industrial level. Elliott (2001) from the industry level, the relationship between advertising and sales experience analyses, “he thought, advertising and marketing industry level data can reflect the relationship between intra-industry competition relations, to test whether there is a stable equilibrium between advertising and sales relationship for a long time, he used the food industry and the soft drink industry, 1983-1992 quarterly data to examine the relationship between advertising and sales, the results showed that the food industry in the market is not saturated state advertising and consumption there is a cointegration relationship, advertising sales have important significant stimulation effect; For the soft drinks market, and its competitive advertising and consumption to long-term stable cointegration relationship exists, which suggests that in the oligopoly market, the market share of the expenses of advertising is to industry redistribution, or advertising spending just to offset or hinder the competitor’s advertising budget, so advertising and sales are not there is a cointegration relationship. Similar to Elliott’s conclusion, Cavaliere et al. (2001)’s study on the Italian beverage industry also found the same conclusion, that there is no significant stimulation effect of alcohol advertising on sales.

However, most empirical studies have come to the opposite conclusion. Numerous studies have found that advertising has a significant impact on sales, which can be long-term or short-term (Landes&Rosenfield,1994; Pa ton, 20 02). Yiannaka et al. (2002) examined the surface data of 34 meat enterprises in Greece from 1983 to 1997 and found that the total advertising expenditure had a significant impact on meat sales. Studies on non-durable goods industry by Franses(1994) and Donovan et al. (2000) also found a long-term stable relationship between advertising and sales. Esteve, etc. (2006) extends the previous studies, they use the British car industry in 1971-2001 annual time series data, adopting multiple structure mutation testing method for car segment of cointegration test, the results show that the six segments in the advertising and sales, there is a long-term equilibrium relationship, but there are four market relations unstable; Structural shocks occurred mainly in the late 1970s and early 1990s, in line with the two recessions; If the impact of external structural changes is not taken into account, the stimulus elasticity of advertising to sales is in a declining state.

In a recession, companies typically spend less on research and development and advertising to maintain liquidity. In this situation, the relationship between advertising and sales are controversial, Kamber (2002) study that AD spending would increase the company sales and profits, but there are also scholars (Kijewski, 1982) think cut back our advertising budget will not bring profit decline, Lamey (2007) argues that during the recession, consumers will be from advertising spending levels higher national brands to lower prices of goods brand, thus reducing advertising sales results. Srinivasan(2011) studied the surface data of more than 10,000 enterprises (1969-2008), and the results showed that the advertising effect during the recession was related to the market share, financial status and product types of enterprises, among which 92% of B2C enterprises had excessive advertising. Bruce et al. (2012) studied that artificial advertising can affect brand perception, such as brand recognition, emotion and experience, so as to stimulate sales. Charles(2014) et al. took the fast food brand McDonald’s as the research object, and found that the relationship between advertising quality and sales growth was significant based on the consumer interview data.

There is also debate about the causal relationship between advertising spending and sales. Based on Australian data, the research of Metwally et al. (1981) confirmed Galbraith’s hypothesis that advertising intensity has a direct and obvious stimulating effect on sales. Zanias (1994) found that there was a long-term equilibrium relationship between advertising expenditure and sales, and the error correction model showed the convergence effect of advertising expenditure and the obvious short-term effect of advertising on sales. But the company only makes drugs for women, and other products that extend to the corporate level need further testing. At the same time, the co-integration relationship between advertising and sales at the enterprise level does not mean that there is a co-integration relationship at the industrial level. Theoretically, the co-integration relationship between advertising and sales only exists at the level of product brand, and is a product with relatively obvious advertising promotion effect. Mouza et al. (2007) used VAR model to test the monthly data of 10 Greek automobile brands.

The model showed that the advertising expenditure increased by 10% and the expected sales growth was 2.5%, which was similar to the trend in reality, but the model did not support the stimulative effect of sales on advertising. Marattin (2008) used the quarterly data of advertising and consumption in Italy from 1980 to 2000 to establish the autoregressive distribution lag model for testing, and the results showed that advertising expenditure had a significant positive impact on consumption. Contrary to the conclusion that advertising stimulates consumption, many scholars (Abe,1995) believe that the stimulating effect of sales on advertising is greater than that of advertising on consumption. In addition, Darrat(2002) used the annual data of the United States from 1948 to 1995 to conduct an empirical test on the relationship between advertising and sales. The results did not verify Galbraith’s hypothesis, but drew the conclusion that sales stimulate advertising, which made a reasonable explanation from the perspective of signal theory. There are few studies on the Chinese market. Ouyang et al. (2002) studied the Chinese market of household durable consumer goods such as color TV sets, refrigerators, washing machines and microwave ovens, and believed that advertising has a long-term effect on sales. Zhou et al. (2003) also found in their study of the Chinese market that advertising has a long-term effect on household durable consumer goods, while it does not have a long-term effect on fast moving consumer goods.