This article is from WeChat public account:Internet and Entertainment Strange Group (ID:TMTphantom), author: head of the monster group Pei Pei, from the title figure: vision China
Since 2018, the so-called “lipstick effect” has been widely circulated in the media entertainment industry and investment circle: consumers will consume more entertainment content when the economy slows down and the macro environment is unstable. (movies, dramas, games, anime, etc.), which brings a thriving “counter-cyclical market” to the media and entertainment industry.
However, this view has been ruthlessly denied by the reality – from January to October 2019, the national box office has only increased by 6% year-on-year, which is still in the “Wandering Earth”, “Avenger League 4”, “Where”, etc. In the case of God’s blessing; the national game market has only a single-digit growth in the same period, of which the mobile game market may grow slightly over 10%, but most of the growth was taken away by Tencent.
What the hell is going on? Has the “lipstick effect” expired?
In fact, the “lipstick effect” has never existed – whether in China or in the United States. If you’ve seen movies from the Hollywood Golden Age (1920-1940s), you’ll know how the Great Depression left a miserable impression on American filmmakers. The so-called “selling sales of lipsticks after the 9/11 incident” has never been supported by data. To put it plainly, it is an urban legend.
Here, the thief group used a large amount of historical data to conduct a regression analysis of the film and game industries in the United States and China, and completely falsified the “lipstick effect.” Of course, we also want to include the episodes, live broadcasts, animation, music and other industries in the regression analysis. Unfortunately, these industries do not have authoritative and reliable sales data, and can only be reserved for future opportunities.
Everything in the world, I am afraid of the word “serious”. Obviously, most people who used to believe in the “lipstick effect”I have not done any empirical research; therefore, this strange thief has done it. Believing in science, believing in data, and believing in logic are the consistent principles of this strange group. Two plus two is always equal to four, not occasionally equal to four; practice is the sole criterion for testing truth, not one of multiple standards. Let us use the scientific method to pronounce the death penalty of the “lipstick effect”.
The economic history data of the United States does not support the existence of the “lipstick effect”
When investors and the media mentioned the “lipstick effect” of the media and entertainment industry, the most common case was the American film industry – there is a saying that in the 1929 and 2008 economic crises, Hollywood was “blessed by misfortune”. Have a good time. Is this statement true? Unfortunately, after analyzing the data of recent decades, we conclude that no matter whether there has been a “lipstick effect” in history, at least we have not observed this phenomenon since the 1980s.
From the annual data, American movies and games are strong cyclical industries
First, let’s take a look at the general statistics course. To judge the correlation between two things, regression analysis is the most common tool: one data is the dependent variable and the other is the independent variable. If there is a clear linear relationship, it can be said that the two are related. Here, we take the US box office income (and the income of the US game industry) as the dependent variable, and take the US macroeconomic data (including GDP, resident disposable income, personal consumption expenditure, unemployment rate, etc.) as the independent variable. It must be noted that all of our raw data is based on nominal values and has not been adjusted for inflation. Since both the independent variable and the dependent variable contain inflation factors, they will self-compensate during the regression analysis.
In the regression analysis, three indicators are very important. The first is the slope, which embodies the strength of the linear relationship between the two variables. For example, if Y = 0.56X + 16, it means that each time X changes by one unit, Y will change by 0.56 units; the smaller the slope, the less important the linear relationship will be.
The second one is R-Square, which embodies the ability of the independent variable X to interpret the dependent variable Y. For example, if we find that R-Square = 0.96, it means that 96% of the change in Y can be explained by the change of X. The linear relationship between the two is perfect and persuasive.
The third is P-Value, which embodies the statistical significance of linear relationships. The smaller the P-Value, the linearity between the independent variable and the dependent variableThe greater the likelihood of a relationship. Academically, P-Value generally needs less than 0.05 to make sense; in practice, it can be appropriately relaxed.
The US box office data is very well preserved. We used all the annual data since 1980 for regression analysis. The conclusion is that the US film industry is a strong cyclical industry, macroeconomics. The correlation coefficient between data and movie box office income is very large, the explanatory power is very strong, and the statistical significance is significant. Similar conclusions can be drawn whether using GDP, disposable income of residents, or total expenditure of individuals as independent variables.
What’s more, regardless of the current or previous macro data, the conclusions of the regression analysis are the same; this shows that macro data such as GDP not only affects the box office of the year, but also the box office of the next year’s movie. Forward-looking. In fact, even with the naked eye, it can be seen that the data points defined by the US box office income and GDP or disposable income of residents are almost on a perfect line, which is the linear relationship that statisticians dream of.
The US authoritative statistics for the game industry (including console games, computer games, and mobile games) began in 1996 with a total of data. The amount is significantly less than the film industry. However, our conclusions have not been affected – there is a strong relationship between the operating income of the US game industry and GDP, disposable income of residents and total personal expenses., a significant linear relationship. Moreover, the slope of the game industry’s revenues for macroeconomic data is significantly larger. In other words, every change in macro data such as GDP has a greater impact on the gaming industry than on the film industry.
In this case, can we announce that the US film and game industry is a strong cyclical industry, and the “lipstick effect” has been falsified? And slow! It should be noted that since the 1980s, the United States has been one of the fastest growing economies in the developed countries. Almost every year, GDP has increased, while movie box office revenues and gaming industry revenues have also increased. This is easy to cause “false positives” – the two sets of data just change in the same direction, not necessarily causal. In addition, the investor’s timeline may not be as long as the year, or it may be quarterly. We also need to conduct a more in-depth analysis of quarterly data, although quarterly data may not be as reliable as the annual.
From the quarterly data, American movies and games are non-cyclical industries
Fortunately, with BoxOfficeMojo, we get continuous US film industry quarterly data. Once the quarterly data is exchanged, the linear relationship between movie box office income and macro indicators such as GDP and disposable income disappears – R-Square has plummeted from around 0.95 to around 0.01, and has lost any persuasive power; P-Value has risen sharply. The statistical significance is not established at all. Whether using the current indicator, or the first one or the first one, the conclusions are similar. In other words, the US film industry is non-periodic, calculated on a quarterly basis.
However, we still can’t think of the “lipstick effect”: the so-called “lipstick effect” means that during the economic downturn, consumers will invest more in entertainment such as movies, which is “counter-cyclical”; Based on statistical analysis of quarterly data, only “aperiodic” can be derived instead of “anti-cyclical”. If investors only want to find an industry that is “unrelated” or “very weak” to macroeconomic indicators, they have a lot of choices and don’t have to rely on the film industry.
The quarterly data of the US game industry is hard to find, and the continuity and authority of common third-party databases are insufficient. Therefore, we can only use the quarterly revenue data of four large listed game companies: Activision Blizzard, EA, Take-Two and Zynga; among them, Activision Blizzard and EA’s revenue scale is significantly larger. It must be acknowledged that the use of the data from the four companies mentioned above is not rigorous, as they all have large revenues from outside the United States and must also consider differences in accounting policies. However, this is already the most reliable quarterly data proxy variable we can imagine.
After using the above proxy variables, our conclusion is: on a quarterly basis, there is no linear relationship between the operating income of the US game industry and macroeconomic indicators such as GDP and disposable income, and statistical significance is almost non-existent. This is similar to our conclusions about the American film industry, with one exception – the average hourly wage in society.
Our statistical analysis shows that there is a linear relationship between the quarterly income of the US game industry and the average hourly wage of the society. The P-Value is only 0.02, reaching a statistically significant level. Unfortunately, R-Square is only around 0.1, which means that only 10% of the game industry’s revenue changes are determined by changes in the average hourly wage. This seems to be common sense: whenever the average hourly wage increases, consumers have more money to buy the game, but only a small part of it is really used to buy the game.
Based on data, research case: What is the cyclicality of the entertainment industry?
From the annual data, the US film and game industry is highly cyclical; from the quarterly data, they are all non-periodic. That’s it? Can we still see a little deeper? Data analysis itself can only involve the extension of things, not its core; it can only tell us “how things change”, and can not tell us “why things change.” To thoroughly understand the cyclical problems of the US entertainment industry, case studies must be conducted.
In the past three decades, there have been three economic recessions in the United States: 1991, 2002, and 2008-09. We can see that in the 1991 recession, the film industry showed a strong cyclicality, with box office revenue and GDP basically changing in the same direction; in 2002, the film industry reflected counter-cyclicality, when GDP was the worst. It was the time when the box office grew at a faster rate; in 2009, the box office revenue rose first against the GDP trend, and then fell against the GDP trend, which seems to reflect a strong counter-cyclicality. If you only look at the last two recessions, you may indeed come to the conclusion that the “lipstick effect” exists. Why is the situation in the three recession periods so different?
Smart readers have probably guessed the answer: product cycle! In the second quarter of 2002, the United States released two super IP blockbusters: “Spider-Man” and “Star Wars Prequel 2”, which prompted the US box office to recover before GDP. As usual, this volume of movies is best suited for the summer season in the third quarter; if the two films are really released during the summer, we will see a good correlation between US box office revenue and GDP, both At the same time a strong recovery.
In the second quarter of 2009, under the impetus of “Transformers 2” and “Wolverine”, the US box office continued to grow year-on-year; in the fourth quarter of that year, box office revenues again showed rapid growth, this time almost entirely “Avatar” Contribution. Works such as “Avatar” that have made great breakthroughs in technology, no matter which year they are released, will completely change the electricity of the year.Film market. Since then, 2010 has happened to be the low tide of Hollywood blockbusters. It is no wonder that the box office receipts fell against GDP.
Our conclusion is simple: the US film industry is affected by both the economic cycle and the product cycle. On a quarterly basis, the product cycle is at the top of the list, because no production company or distribution company can guarantee that the work will be released in a certain quarter, and the commercial blockbuster will be delayed or filed for several months. On an annual basis, the product cycle is largely “smoothed” and the economic cycle is at the top of the list because it determines the real needs of the audience. Therefore, it is logical for the US film industry to exhibit strong periodicity in the year and non-periodicity in the quarter.
The situation in the US game industry is another story: in the 1991 recession, the industry was almost non-existent, and its data was meaningless; in 2002, game industry revenues rebounded earlier than GDP, and the rebound momentum was fierce. It reflects the counter-cyclicality; by 2008, the income of the game industry has changed almost in the same direction as GDP, reflecting a strong cyclicality. This can not help but make people think: With the expansion of income scale, is the US game industry moving from “non-cyclical” to “cyclical”? In the next US recession (possibly this year or next year), will the quarterly revenue of the gaming industry show a strong cyclicality? The possibility exists, and it is not small.
In addition, the proxy variables we selected for the US game industry’s quarterly revenue include a certain amount of income from outside the US, which also weakens the correlation between the data itself and US macro data. We believe that the cyclicality of the quarterly data for the gaming industry is likely to rise if it can completely exclude revenues from markets outside the US; however, we have not been able to confirm this speculation.
What is the situation in China? Movies and games need to be separated separately
The historical data of the United States can give us a lot of inspiration, but what investors care most about is the Chinese market. Although the historical data of the Chinese film and game industry has accumulated less,It is always based on historical data. After conducting a similar statistical analysis, we found that the Chinese film industry is non-periodic on a quarterly basis, but the game industry is cyclical. In any case, there is no counter-cyclicality in any of them, that is, there is no “lipstick effect.”
Chinese film industry: non-periodic? The key is too many disturbance factors
Taiwan and the State Administration of Radio, Film and Television, we can know the quarterly data of box office receipts of Chinese movies since 2007. In fact, we can go back to longer, but it doesn’t make much sense – the film industry and economic environment before 2007 are very different from today, and the comparability is very weak. Regression analysis of movie box office income and macro indicators such as GDP, disposable income, and PMI (all do not exclude inflation), we found that almost all linear relationships are not established, and the Chinese film industry is a non-cyclical industry with only one indicator. Exception: unemployment rate.
Yes, regression analysis shows that the positive correlation between the growth rate of Chinese box office receipts and the urban population unemployment rate is statistically significant, P-Value is only 0.01; the higher the unemployment rate, the faster the box office income growth! However, R-Square is only 0.13, which means that only 13% of the movie box office growth is caused by rising unemployment. We do not believe this linear relationship – China’s urban unemployment rate coverage is not enough, the historical range of change is very small, it is likely to lead to “false positives.” In any case, on a quarterly basis, we still believe that the Chinese film industry is a non-cyclical industry.
As we mentioned above: “non-periodic” does not mean “counter-cyclical”, the latter is the “lipstick effect.” In fact, the non-periodicity of the Chinese film industry is easy to explain, that is, there are too many disturbance factors, including but not limited to the product cycle. In 2015, with the product cycle going up and the influx of capital, the box office receipts recorded a staggering 49% increase. In 2016, due to the reduction in ticket replenishment and the SARFT’s clear fight against “fake box office”, the box office revenue growth rate Suddenly shrinking. In the US, the movie marketSupply is affected by the product cycle, and demand is affected by the economic cycle; in China, even the demand is affected by a series of complicated factors such as ticketing and new theaters.
From the quarterly comparison between movie box office revenue and nominal GDP growth, we can also intuitively feel that the film industry and GDP sometimes change in the same direction, sometimes in reverse, and there is no rule to follow. At the beginning of 2009 and early 2012, the nominal GDP growth rate suddenly slowed down twice, and the movie box office growth rate was basically stable. In 2014-16, the nominal GDP growth rate slowed down and bottomed out. During the period, the box office growth rate appeared many times. Variety. All in all, judging the popularity of the film market through the economic cycle is an inefficient and unworthy way to learn.
Chinese game industry: has strong periodicity, so be careful!
We have two ways to access historical data from the Chinese game industry: iResearch and the China Game Industry Committee (Gamma Data). Among them, the data of iResearch is quarterly, and the China Game Industry Committee is semi-annual. In practice, we found that the conclusions of using these two sets of data are not much different; moreover, since 2015, the correlation between the two sets of data has exceeded 80%, which can be counted as a set of data. Therefore, we decided to use iResearch’s quarterly data since 2010.
Regression analysis shows that there is a linear relationship between China’s game industry income and GDP, per capita GDP, per capita disposable income, and urban residents’ consumption expenditure on a quarterly basis. Statistical significance is high, P-Value Even close to 0 – this shows that our conclusion is almost no mistake. From the perspective of R-Square, the situation is a bit worse: only about 50% of the change in operating income in the gaming industry can be explained as macroeconomic changes, but this ability to explain is already very good. In other words, the Chinese game industry is a strong cyclical industry!
The macro indicator that has the strongest ability to explain the income of the game industry is the per capita disposable income of urban residents. It can be seen from the comparison of quarterly data that since 2007, the game industry’s income and per capita disposable income (both excluding inflation) have changed in the same direction in most cases. There have been only two exceptions: in the fourth quarter of 2014 and the second quarter of 2018, the per capita disposable income and the income of the game industry did not change in the same direction, and the latter was affected by the suspension of the game version. Logically, the rise and fall of disposable income does affect the game’s propensity to consume.
Why is the Chinese film industry non-periodic and the game industry highly cyclical? We think there are at least three reasons. First of all, the scale of the Chinese game industry is very large; even according to the most conservative estimates, the game market in 2018 is more than 200 billion yuan, more than three times the size of the film market. Such a large industry cannot be without periodicity. Secondly, the number of products supplied by the game industry is much larger than that of the film industry. In 2017, more than 9,300 games in China were approved, but only more than 700 movies were released, so the game industry was slightly less affected by the product cycle. Third, the game industry has suffered from fewer external disturbances in the past few years. Although there are also water-spraying phenomena, the impact is not large compared to the ticketing of the film industry.
By the extensive statistical analysis and case analysis above, I believe readers have already understood our conclusion: empirically, it is impossible to prove that there is a “lipstick effect” in the media and entertainment industry represented by movies and games, and in the United States and China. This is the case. Of course, the media and entertainment industry not only includes movies and games, but also TV dramas, online videos, short videos, live broadcasts, animations, novels…