网络消费者信心与宏观经济波动关系研究

发布时间：2018-01-07 16:18

本文关键词：网络消费者信心与宏观经济波动关系研究　出处：《吉林大学》2017年博士论文　论文类型：学位论文

【摘要】：网络搜索相关研究的开展得益于互联网的普及与繁荣,它是随着数据挖掘技术的发展而兴起的大数据研究。最开始搜索引擎中的数据被应用于流行病研究领域,随后,许多学者将这一技术应用到失业率、电影票房、商品销量等其他方面的分析和预测,这些研究充分证明了搜索引擎中的搜索数据与经济行为之间存在相关性。文章将网络搜索数据看成是消费者需求在网络上的映射,构建出基于网络搜索数据的网络消费者信心指数(Web-query-based Consumer Confidence Index,简称WCCI),研究其与宏观经济的关系,并试图论证该指数在即时性、预测准确性等方面的优点。多角度、多层次的实证分析可以证明,网络消费者信心指数可以作为众多经济景气指数的补充来判断宏观经济形势。为了论证网络消费者信心与宏观经济之间的关联性,论文主要进行了以下几个方面的研究:一、构建网络消费者信心指数。消费者信心指数是一个具有反映消费者消费趋向、预测经济运行功能的先行指标。在我国,这一指标由国家统计局负责监测和公布,指标的公布通常具有一个半月左右的时滞。而网络用户的需求以及意愿等信息可以通过对网络搜索数据的整理和分析得以反映,这些互联网络的使用者同时也是构成市场主体的生产者和消费者,他们供需关注重点的变化在市场中表现为商品或服务的数量及价格的波动,体现在互联网上则为搜索、浏览等行为的变化即搜索数据的变化。因此,搜索行为与市场主体的经济活动之间是有联系的。此外,与传统消费者信心指数的数据获取途径相比,网络搜索技术可以通过获得数据总体的方式避免抽样调查的局限性,而且数据量更大,具有更容易获得的特点。我们引入即时变化的网络搜索数据构建网络消费者信心指数,该指数能将网络消费者信心受外部因素冲击的影响情况实时地反映出来,我们可以利用网络消费者信心指数进行一些具有时效性问题的研究。在介绍了我国现有消费者信心指数的编制、公布情况后,我们将消费者信心指数的时滞性作为切入点,探索网络搜索数据对消费者信心指数进行补充的可行性。在所有类别的商品或服务中,消费者在购买某些特殊类别的商品前会更倾向于使用网络搜寻信息辅助决策,而在构建网络消费者信心指数时,我们选取的各类别关键词基本涵盖了这些类别的商品或服务,因此数据来源具有广泛性,指标获取具有实时性。网络消费者信心指数构建完成后,我们主要论证了以下观点:不同行业类别的商品或需求对于互联网的依赖程度和响应程度不同,其动态波动情况和周期性变化特点各不相同,网络搜索数据与消费者信心和消费需求之间具有一定的相关性。二、货币政策、财政政策等宏观经济政策对于一个国家的宏观经济稳定发展有着不言而喻的重要作用。宏观经济政策效果及其影响因素是政府和学者都十分关注的问题。通常情况下,货币政策的主要目标是调节通胀、稳定物价,而财政政策则用于扩大内需和治理萧条。信心是宏观经济政策传导到微观经济主体这一过程中十分关键的环节。信心水平的高低对于宏观经济政策效果的好坏将产生一定影响。在我国经济进入中高速发展的新常态形势下,从网络消费者信心这一独特视角分析宏观经济政策的传导机制和作用效果或将为政策制定、政策实施及组合等提供新的思路。在我们的研究中,采用马尔科夫区制转移向量自回归模型(MSVAR)对网络消费者信心与财政政策、货币政策之间的关系和相互作用进行研究,并探索变量的动态转移机制和波动情况,得出了如下结论:网络消费者信心和宏观经济政策及其作用效果的动态变化存在区制转移的特征,我国的网络消费者信心可以分为高信心水平区间和低信心水平区间,且区间的划分与经济政策的扩张和收缩是相对应的。不同政策工具与网络消费者信心的关联性强弱也因工具性质的不同有所差异。网络消费者信心作用于宏观经济政策效果的程度并不是稳定不变的,而是有规律地变化:这种影响随着信心的增强而增强,随着信心的减弱而减弱。网络消费者信心与宏观经济政策之间的关联性得到了论证。三、物价波动是宏观经济波动的重要组成部分,同时也关系到人们日常生活的方方面面,而网络消费者信心作为消费需求在互联网上的映射,可能与物价波动存在密切联系。在这部分研究中,我们以我国居民消费价格指数(Consumer Price Index,简称CPI)作为测度物价波动的量化指标,以网络消费者信心指数来度量网络消费者信心,识别和分析我国网络消费者信心和物价波动动态过程中的内生转移机制。实证分析结果表明,结构性突变确实显著存在于网络消费者信心指数增长率和CPI增长率序列中,同时,研究结果也表明网络消费者信心指数增长率的波动性要强于后者。从长期发展的视角来看,CPI相比较网络消费者信心指数而言,更容易保持低速增长。而对于网络消费者信心指数来说,增长率与波动性之间可能存在正相关关系,即网络消费者信心指数增长率越大则波动性可能越强,网络消费者信心指数增长率越小则波动性可能越弱。以我国的网络消费者信心和物价波动为研究对象,探索网络消费者信心指数增长率与居民消费价格指数增长率序列中可能存在的结构性突变,意图通过马尔科夫区制转移模型得到更精确的估计结果,进而证明网络消费者信心与物价波动之间的紧密联系。对两个时间序列内生转移机制的识别过程,为我们论证网络消费者信心和物价波动相互作用情况提供了新的视角,网络消费者信心指数的高波动性和快速增长等特点有利于经济拐点的判断和通货膨胀情况的监控。四、经济增长也是宏观经济波动的重要组成部分,而经济增长的主要构成因素即为投资、消费和净出口。工业增加值(简称IP)可以较好地衡量投资所代表的经济增长水平,为了探索其与网络消费者信心的关系,我们利用动态条件相关多元模型(DCC-GARCH)、结构向量自回归模型(SVAR)对两者的动态特征进行定性和定量研究,通过统计特征分析、平稳性分析、格兰杰因果检验和协整检验等方法对两个序列进行了纵向比较,初步证实了网络消费者信心与工业经济增长之间的相关性,网络消费者信心指数可以格兰杰引起工业增加值的变化;而DCC-GARCH模型和SVAR模型则证明了网络消费者信心指数与工业增加值之间存在显著的动态相关关系,且相关关系波动性较大,网络消费者信心指数可以解释约20%的工业增加值波动。消费者通过网络搜索而获得的信息会影响他们的微观消费需求和购买决策,这种变化反映在宏观层面则为总需求的波动,总需求波动会影响总供给水平,进而使整个宏观经济产生波动;供给方面,经济增长会作用于生产行为,同时改变消费者可支配收入,人们的消费预期就会发生变化。网络消费者信心与工业经济增长之间相互影响、相互作用。在分析动态相关系数波动和网络消费者信心对经济增长的冲击情况时,文章从家庭可支配收入的角度进行了论证。作者认为,可能是宏观景气程度影响了动态相关系数的变化:当宏观经济比较景气时,消费增加对投资的“挤占效应”比较小或者被抵消,经济增长速度加快;而当经济比较低迷时,消费需求增加所导致的消费增加对于投资的“挤占效应”比较明显,经济增长放缓。这种假设在动态相关系数和宏观预警指数序列的对比中得到了证实。五、经济增长的相关影响因素分析对于把握宏观经济有至关重要的作用,只有掌握了相关因素的变化规律,才可以让经济主体更好地把握经济发展趋势,并让决策制定者提前制定调控政策,促进市场资源的合理化配置。国内生产总值(Gross Domestic Product,简称GDP)作为一个反映宏观经济的总量指标,它所具有的衡量经济增长水平的作用一直被学术界认可。根据支出法核算的GDP,主要由投资、消费和净出口等三个部分组成。在我国国家统计局公布的经济数据中,GDP的增长率、累计值等相关数据都是按季度公布,通常有一个半月左右的时滞;与其他月度经济数据相比,频率差异和时滞差异可能会使各个经济主体对于当下宏观经济的总体状态和趋势无法做出准确判断。传统的线性回归模型在构建过程中需要频率相同的时间序列来进行分析,不同频率的数据需要转换成相同频率的数据再进行建模。频率转换的过程中会有信息的损耗和丢失,如何利用不同频率的数据直接构建模型进行分析,最大程度上保留数据中的有效信息是近年来学术研究的热点之一。网络搜索数据与宏观经济密切相关,利用高频的网络搜索数据研究其与宏观经济变量的关系是合理的。本文运用混频数据模型研究网络消费者信心与经济增长的相关关系。首先建立起包含GDP自身滞后期的向量自回归预测模型作为基准模型,而后分别建立两个不同多项式表达的混频预测模型,就这三个模型的残差、预测指标等参数做出综合比较,结果表明,加入了高频的月度网络消费者信心指数后,模型的拟合程度及预测效果与传统自回归模型相比明显提高,混频数据抽样模型在研究网络消费者信心与经济增长关联性过程中的优越性得到体现。我们利用混频数据模型解决了高频数据转换成低频数据过程中的信息损失问题,最大程度上保留了数据信息,从而证明了网络消费者信心指数对于预测我国经济增长的重要作用。借助网络搜索数据,我们依据相关理论构建了网络消费者信心指数,而后分别研究了该指数与物价波动、经济增长及宏观经济政策之间的相关性。各个实证结果充分证明,网络消费者信心指数能够较好地反映我国消费者的信心水平情况,它与各个宏观经济变量之间存在着相互影响与相互作用的关系,与经济主体之间存在着较为密切的联系。与此同时,网络消费者信心指数也可以作为信心的代理变量,它有效地印证了信心在经济主体的各种经济行为中不可忽视的重要作用。总而言之,网络消费者信心指数基本可以作为一个反映我国宏观经济的补充指标,为宏观经济预警、经济政策的制定、经济拐点的预判等方面提供新的思路。
[Abstract]:Web search related research popularity and prosperity thanks to the Internet, it is the study of big data emerges with the development of data mining technology. The beginning of the search engine data is applied to epidemiological research field, then, many scholars have applied this technique to the unemployment rate, the film at the box office, the analysis and prediction of commodity sales the other hand, these studies proved that there is a correlation between the search data and economic behavior in the search engine. The Internet search data is mapped as consumer demand on the network, building a network search network based on the data of the consumer confidence index (Web-query-based Consumer Confidence Index, referred to as WCCI), to study the relationship with macro the economy, and attempts to prove the index in real time, accuracy and other aspects of the advantages of multi angle, multi-level empirical analysis can permit Next, the network consumer confidence index can be regarded as a supplement to the many economic boom index to judge the macro economic situation. In order to demonstrate the relationship between network consumer confidence and macro economy, the paper mainly studied the following aspects: first, constructing the network consumer confidence index. Consumer confidence index is a reflection of consumer consumption trends, in advance the economic operation index prediction function. In our country, the indicators for monitoring and published by the National Bureau of statistics, the index released usually has a delay of about one and a half. While network users desire and demand of information through the collation and analysis of data on the network search to reflect, these Internet users at the same time a producer and consumer market, they focus on the changes of supply and demand in the market for goods or services number Volume and price volatility, reflected in the Internet for searching, browsing behavior changes that search data changes. Therefore, there is a link between the search behavior and the main body of the market economy. In addition, the consumer confidence index and the traditional ways of data acquisition compared to network search technology can obtain the overall data the way to avoid the limitation of sampling survey, and the data quantity is large, with easier access to the data network construction. The consumer confidence index search we introduce an instant change in network, the network consumer confidence index can be affected by external factors impact the real-time effect reflected, we can use the Internet for some of the consumer confidence index the issue of timeliness research. Introduces the preparation of China's existing consumer confidence index, announced the situation, we will be the consumer confidence index Time delay as a starting point, to explore the feasibility of Internet search data on consumer confidence index was added. In all categories of goods or services, consumers will be more inclined to use the Internet to search for information in the purchase decision of some special categories of goods, and in the construction of network consumer confidence index, we selected the various basic keywords don't covers these categories of goods or services, so the data source is wide, obtain the real-time network index. The index of consumer confidence after the construction, we mainly discusses the following points of view: different industry categories of goods or the demand degree of dependence on the Internet and in response to different degrees, the dynamic fluctuations and periodic variation characteristics different web search between data and consumer confidence and consumer demand has a certain correlation. Two, monetary policy, fiscal Play an important role in the macro economy it is self-evident for a country's stable development policy and other macroeconomic policies. The factors of macroeconomic policy effects and the influence of government and scholars are very concerned about the issue. Usually, the main objective of monetary policy is to regulate inflation, price stability, and fiscal policy to expand domestic demand and treatment of depression. Confidence is the key macroeconomic policy conduction to the micro economic subject in this process. The confidence level is good or bad for the effect of macroeconomic policy will have a certain impact. In the rapid development of China's economy has entered a new normal situation, from the unique perspective of network consumer confidence analysis and the mechanism of the effect effect the macro economic policy or for policy formulation, policy implementation and provide new ideas and combination. In our study, by Markoff The transfer system of vector autoregressive model (MSVAR) on the confidence of network consumers and fiscal policy, monetary policy and the relationship between the interaction of variables and explore the dynamic transfer mechanism and fluctuation, draw the following conclusion: the dynamic changes of the network consumer confidence and macroeconomic policy and the effect of the characteristics of the transfer system the network of consumer confidence in China can be divided into a high level of confidence interval and the low level of confidence interval, expansion and contraction and interval division and economic policy is the corresponding. Relationships between different policy tools and network consumer confidence but also because of the different nature of the tool. Network consumer confidence in effect the degree of macroeconomic policy is not stable, but changes: this effect is enhanced with increased confidence, with the weakening of confidence Weakened. Connection between consumer confidence and macroeconomic policies have been demonstrated. Three, the price fluctuation is an important part of macroeconomic fluctuations, but also related to all aspects of daily life, and network consumer confidence as a consumer demand on the Internet map, might have close links with the price fluctuation. This part of the study, we in China's consumer price index (Consumer Price Index, referred to as CPI) as a quantitative index to measure the price fluctuations, the network consumer confidence index to measure consumer confidence, identification and analysis of the transfer mechanism of China's network of consumer confidence and price fluctuation in the dynamic process. The results of the empirical analysis show that structural mutations significantly exist in the network consumer confidence index growth rate and CPI growth rate series, at the same time, the results also show that the network. Stronger volatility consumer confidence index growth rate than the latter. From the long-term development perspective, compared to CPI network consumer confidence index, more likely to maintain a low growth. As for the network consumer confidence index, there may be a positive correlation between the growth rate and volatility, namely the network consumer confidence index growth rate is high volatility may be more strong, the network consumer confidence index growth rate the smaller the volatility may be weaker. In our network of consumer confidence and price fluctuation as the research object, to explore the network consumer confidence index growth rate and the consumer price index growth rate possible structural breaks in the sequence, intent to estimate more accurate through the Markov regime switching model, and prove the close relationship between network consumer confidence and price fluctuation. The two time series in The identification process of transfer mechanism, provides a new perspective for us to demonstrate the network consumer confidence and the price fluctuation in the interaction network, the consumer confidence index high volatility and the rapid growth of the characteristics of monitoring judgment and inflation to economic inflection point. Four, an important part of economic growth and macroeconomic fluctuations, and the main a factor of economic growth is investment, consumption and net exports. The industrial added value (IP) can be used to measure the investment on behalf of the level of economic growth, in order to explore its relationship with the network of consumer confidence, we use dynamic correlation multivariate model (DCC-GARCH), the structural vector autoregression model (SVAR) and qualitative the quantitative study of the dynamic character of the two, through statistical analysis, stability analysis, Grainger Granger causality test and cointegration test method for two sequences The longitudinal comparison, confirmed the correlation between network consumer confidence and economic growth, the network consumer confidence index can cause changes in the Grainger industrial added value; and the DCC-GARCH model and the SVAR model proved that the network consumer confidence index and the industrial added value between the dynamic correlation, and the correlation between fluctuation the network, the consumer confidence index can explain about 20% of the industrial added value fluctuation. Consumers through the Internet search and access to information will affect the micro consumption demand and their purchase decision, this change is reflected in the macro level, the total demand for the fluctuation of total demand fluctuation will affect the total supply level, thus the whole macro economy fluctuation; the supply side, the economic growth will effect on the production and changing consumer behavior, disposable income, people's consumption is expected Change. Network between consumer confidence and industrial economic growth mutual influence and interaction. In the analysis of the dynamic correlation coefficient and fluctuation of network consumer confidence in economic growth impact, the family disposable income angle are discussed. The author thinks that can change the degree of macroeconomic dynamic correlation coefficient: when the macro economy economy, increased spending on investment "crowding out effect" is very small or have been offset by faster economic growth; and when the economic downturn, the increase in consumer demand caused by the consumer to increase investment in the "crowding out effect" is more obvious, the slowdown in economic growth. The comparison of dynamic correlation coefficient and the macro warning index sequence has been confirmed. Five, the analysis of the influence factors of economic growth to grasp the macroeconomic is crucial for, Only to master the changes of related factors, it can make the economy subject to better grasp the trend of economic development, and make decision makers to advance the development of regulatory policy, to promote the rational allocation of market resources. The gross domestic product (Gross Domestic Product, referred to as GDP) as a reflection of the total index of macro economy, it has a measure of the level of economic growth effect has been acknowledged by scholars. According to the expenditure method of accounting of GDP, mainly composed of three parts of investment, consumption and net exports etc.. In China's economic data released by the National Bureau of statistics, the growth rate of GDP, the cumulative value of the relevant data are released on a quarterly basis, usually with time delay about a month and a half; compared with other monthly economic data, the frequency difference and delay difference may lead to various economic entities for the overall status and trends of current macro economy cannot make accurate judgment Break the traditional linear regression model. In the process of building the time series the same frequency analysis, different frequency data need to be converted into the same frequency of data modeling. And the loss of loss frequency conversion will have information, how to use different frequency data directly construct model analysis, the maximum extent keep the effective information in the data is one of the hot topic in academic research in recent years. Internet search data is closely related to macroeconomic data search and to study the relationship between macroeconomic variables is the rational use of high frequency network. This paper uses the relevant data model based on network relationship between mixing consumer confidence and economic growth. Firstly, establish the vector including GDP lag Autoregression Prediction Model as the reference model, and then establish two different polynomial expression mixing model, On the residuals of these three models, a comprehensive comparison, to predict the parameters. The results showed that adding high frequency network monthly consumer confidence index, the effect with the traditional fitting degree and the prediction model of autoregressive model is significantly improved compared to the superiority of the mixed data sampling model in the study of network consumer confidence and economic growth related process in the embodiment. We use the data model to solve the high frequency mixing data into information loss of low frequency data process, the maximum extent retained the data, which proves that the network consumer confidence index forecast for economic growth in China

【学位授予单位】：吉林大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：F724.6;F124.8

【相似文献】