Impacts of Digital Economy on Price Stickiness and Monetary Non-neutrality: Evidence from China
Tingfeng Jiang (University of International Business and Economics)
Taoxiong Liu (Tsinghua University)
Ke Tang (Tsinghua University)
January 2021
2020/12/29 Outline o Part 1: Introduction o Part 2: Models o Part 3: Comparison of Online and Offline Price Stickiness o Part 4: Measurement of Monetary Non-neutrality o Part 5: Conclusions and Implications
2020/12/29 2/27 Part 1 Introduction o Motivation ü In recent years, the rapid development of the digital economy has profound impacts on society, especially during the Covid-19 period. ü At the meanwhile, short-term fluctuations in macroeconomics have become more frequent.
2020/12/29 3/27 Motivation p Online prices receive more and more attention. ü BPP(Billion Price Project, http://www.thebillionpricesproject.com/) ü Tsinghua University iCPI (Internet-based Consumer Price Index) program, http://www.bdecon.com/chartsEnglishIndex p Online prices are different from prices in brick- and-mortar stores ü low search costs ü low costs of monitoring competitors’ prices ü low costs of nominal price adjustment
2020/12/29 4/27 Motivation o Key Question 1: Will the rapid development of digital economy affect the micro foundation of short-term macroeconomic analysis? What are the differences between online and offline price stickiness? o Key Question 2: What are the impacts of digital economy on the monetary non-neutrality? What opportunities and challenges does the big data era provide for the analysis of monetary non-neutrality?
2020/12/29 5/27 Literature Review p Basic price adjustment models ü time-dependent pricing model (TDP, Taylor, 1980; Calvo, 1983; Carvalho,2006; Alvarez et al., 2016) ü state-dependent pricing model (SDP, Barro, 1972; Dotsey et al., 1999; Golosov and Lucas, 2007; Nakamura and Steinsson,2010; Alvarez and Lippi,2014) p Theories about the cost of price adjustment ü search cost (Alchian,1969); menu cost(Sheshinski and Weiss, 1977) ü costs of updating information (Mankiw and Reis, 2002) ü observation costs for companies (Alvarez et al., 2011) p Empirical evidence of price stickiness ü Offline prices (Bils and Klenow, 2004; Klenow and Kryvtsov, 2008; Nakamura and Steinsson, 2008/2010) ü Online prices (Cavallo, 2016/2017/2018; Gorodnichenko et al., 2018)
2020/12/29 6/27 Main Findings o We measure and compare online and offline price stickiness with unique data in China, and employs the empirical evidence to calibrate the hybrid heterogeneous multi-sector price adjustment model. o First, we find that the online and offline price change durations are one and half months, and six months, respectively. However, the online and offline absolute price change sizes are similar, namely about 15%. There also exists obvious heterogeneity in different divisions and asymmetry in price increases and decreases. o Second, with current online markets accounting for about 20%, the effectiveness of monetary policy is estimated to be 42%- 53% of the pure offline market. If the online market share reaches 100% in the future, the monetary non-neutrality is only 7% -11% compared to that of pure offline market. 2020/12/29 7/27 Marginal Contributions o Firstly, this paper examines the differences between online and offline price adjustment behavior and its impact on monetary non-neutrality, which is an important problem in the digital economy era. o Secondly, this paper adopts unique online and offline market micro- price data in China. The offline data comes from the Price Monitoring Center of the National Development and Reform Commission; the online data comes from the iCPI program of Tsinghua University. o Finally, this paper shows that impacts of the digital economy on the transmission of monetary policy should not be ignored. In the digital economy era, central banks should pay more attention to high- frequency online inflation indicators, attach more importance to structural monetary policy.
2020/12/29 8/27 Part 2 Models o Hybrid Heterogeneous Multi-sector Price Adjustment Model SDP+TDP ; Refer to Nakamura and Steinsson, 2010 • Hypothesis: ü Dixit-Stiglitz, CES, differentiated goods and labor . ü Sectors s = 1,2, … , S firm j ∈ 0,1 , ∑-/0 ω- = 1 • Household decision: LJQ BLMN70 D ∑U I CJK ∑. 0 EF,CJK >?@ H3{ I/T 5 − -/0 ω- ∫T RS } {BC,DEF,C} 078 04= 0 0 s. t. 9 6 + H Y2 Z ] ≤ ∑. ω ∫ : < RS + Z + ∑. ω ∫ Π RS 3 3 3 3,340 340 -/0 - T -;,3 -;,3 3 -/0 - T -;,3 78 6340 93 :-;,3 = 8 23,340= 5 = <-;,363 63 9340 93 a No-ponzy-game Scheme ^_` 5 ?a40 ≥ 0 a→U a d Transversality Condition ^_` 5 c (?a40)?a40 = 0 a→U 2020/12/29 9/27 Hybrid Heterogeneous Multi-sector Price Adjustment Model SDP+TDP
• Firm decision: U >?@ H3 k (23,34Il-;,34Im hEF,C,iEF,C,jEF,C I/T
l-;,3 = 9-;,3n-;,3 − :-;,3 o-;,3 − 93p-;,3 − q-:-;,3r-;,3 07v v s. t. n-;,3 = u-;,3o-;,3 p-;,3
^w(u-;,3) = x ^w u-;,370 + y-;,3 2 εsj,t~N(0, σs,ε) 0 ÅML Å Intermediate input p Å ÅML -;,3 p-;,3 = [∫T ℎ-;,3 (Ä) RÄ] 7É 9-;,3(ÄÇ F.O.C ℎ-;,3 Ä = p-;,3 93 7É 9-;,3 n-;,3 = n3 93 10/27 Hybrid Heterogeneous Multi-sector Price Adjustment Model SDP+TDP
• Two types of companies: (1) ÑÖ = ÜÖ,á, àÖ. (2) ÑÖ = ÜÖ,â, ä − àÖ • Firm decision of low adjustment cost: U >?@ H3 k (23,34IΠ-;å,34I) {hEFã,C,iEFã,C,jEFã,C} I/T
l-;å,3 = 9-;å,3n-;å,3 − :-;,3 o-;å,3 − 93p-;å,3 − q-,å:-;,3r-;å,3
• Firm decision of high adjustment cost : U >?@ H3 k (23,34Il-;é,34I) {hEFç,C,iEFç,C,jEFç,C} I/T
l-;é,3 = 9-;é,3n-;é,3 − :-;,3 o-;é,3 − 93p-;é,3 − q-,é:-;,3r-;é,3 0 07É 07É 07É • The whole market 9-,3 = (1 − è-)9-;å,3 + è-9-;é,370 11/27 Hybrid Heterogeneous Multi-sector Price Adjustment Model SDP+TDP
• Other conditions 0 0 ü ∑. ∑. n-;,3 = 6-;,3 + -/0 ω- ∫T ℎ-;,3(S)RS , n3 = 63 + -/0 ω- ∫T p-;,3RS
ü <-,3= o-,3 = (1 − è-)o-;å,3 + è-o-;é,3 ó ü Ln>3 = í + ln >370 + î3 where î3~o(0, ïñ ) • Bellman Equation of low adjustment cost firms: 9-;å,370 >3 ô ô 9-;å,3 >340 ò(u-;å,3, , ) = `?@ {l-;å,3 + H3[23,340ò(u-;å,340, , )]} 93 93 {hEFã,C,iEFã,C,jEFã,C} 9340 9340 ô öEFã,C hEFã,C õEF,C õEF,C l-;å,3 = = n-;å,3 − o-;å,3 −p-;å,3 −q-,å( )r-;å,3 hC hC hC hC ô 23,340 23,340 = 93
12/27 Hybrid Heterogeneous Multi-sector Price Adjustment Model SDP+TDP
• Bellman Equation of high adjustment cost firms:
9-;é,370 >3 ô ô 9-;é,3 >340 ò(u-;é,3, , ) = `?@ {l-;é,3 + H3[23,340ò(u-;é,340, , )]} 93 93 {hEFç,C,iEFç,C,jEFç,C} 9340 9340 ô öEFç,C hEFç,C õEF,C õEF,C l-;é,3 = = n-;é,3 − o-;é,3 − p-;é,3 − q-,é( )r-;é,3 (18) hC hC hC hC ô 23,340 23,340 = 93
• Guess and Verify
13/27 Part 3: Comparison of Online and Offline Price Stickiness o Our online data contains prices from more than 100 websites covering the whole basket of Chinese CPI with over 19 million price records, including 8 divisions, 46 groups, and 262 classes. The sample period is from January, 2016 to February, 2019. (iCPI program in China; also refer to Jiang et al., 2020 . o The offline data is from the price monitoring center of the National Development and Reform Commission in China, which includes 126 types of food, daily industrial consumer goods, and services with over 50 thousand price records. The sample period is from January, 2017 to March, 2020.
2020/12/29 14/27 Comparison of overall online and offline price stickiness indicators p The weighted average online and offline price change durations are one and half months, and six months, respectively, with obviously more frequent online price changes. p The online and offline absolute price change sizes are very similar, namely about 15%. p There exists obvious asymmetry in price increases and decreases for online and offline markets.
2020/12/29 15/27 Comparison of offline price stickiness among different divisions p Offline food price adjustments are the most frequent, followed by manufactured consumer goods, and service price adjustments are the least frequent. p Offline services have the largest price change size, followed by food, and manufactured consumer goods have the smallest price change size . p There exists obvious heterogeneity in different divisions and asymmetry in price 2020/12/29increases and decreases. 16/27 Comparison of online price stickiness among different divisions p For the online markets, there is obvious heterogeneity for price stickiness indicators among different divisions. However, price increase frequency is consistently higher than decrease frequency, and price increase size is also higher than decrease size. p For both offline and online markets, the price stickiness indicators of different divisions have obvious heterogeneity and asymmetry. 2020/12/29 17/27 Part 4: Measurement of Monetary Non-neutrality
o Benchmark Parameters
2020/12/29 18/27 Core Parameters Calibration
o We combine the empirical price stickiness indicators
to calibrate the menu cost Ζ- and productivity shock ï.,ù. o We find that, under the same conditions, the menu cost of offline market is significantly higher than that of the online market, but the productivity shocks of different markets have no specific relationship. This indicates that Internet greatly reduces the price adjustment cost.
2020/12/29 19/27 Monetary Non-neutrality based on Online Markets o Monetary Non-neutrality based on Online Markets ü TDP > SDP+TDP > SDP 1.5 3 1 2 ü With Intermediate Input>Without Intermediate Input 1.5 4 ü heterogeneous model> homogeneous model 7 17
2020/12/29 20/27 Monetary Non-neutrality based on Offline Markets o Under the same conditions, monetary non-neutrality based on offline market is significantly higher than that based on online market. (SDP+TDP 1.2017 /0.1378)
2020/12/29 21/27 Online-Offline Markets Combination and Monetary Non-neutrality o We combine the weighted average price stickiness indicators of both online and offline markets to calibrate the model.
2020/12/29 22/27 Online-Offline Markets Combination and Monetary Non-neutrality o According to the current situation, the online market accounts for about 20% in China, and the effectiveness of monetary policy is estimated to be 53% of the pure offline market without intermediate input, which is further reduced to 42% considering the intermediate input. o As the digital economy develops, the online market share will increase, and if it reaches 100% in the future (under an extreme condition), the effectiveness of monetary policy is only 7% to 11% compared to that of pure offline market.
2020/12/29 23/27 Part 5 Conclusions and Implications o Main Conclusions l The price stickiness indicators of online market and offline market in China are not only significantly different, but also have certain commonalities. ü On one hand, the online and offline price change durations are one and half months, and six months, respectively, indicating that online price changes are much more frequent due to low adjustment cost. ü The online and offline absolute price change sizes are very similar, namely about 15%. There exists obvious heterogeneity in different divisions and asymmetry in price increases and decreases. 2020/12/29 24/27 Main Conclusions l Recently, as the proportion of digital economy has been increasing in China, the degree of price stickiness has been significantly reduced (price change frequency has increased significantly), which to some extent has led to a weakening of the effectiveness of monetary policy. ü With current online markets accounting for about 20% in China, the effectiveness of monetary policy is estimated to be 53% of the pure offline market without intermediate input, which is further reduced to 42% considering the intermediate input. ü If the online market share reaches 100% in the future, the effectiveness of monetary policy is only 7% to 11% compared to that of pure offline market.
2020/12/29 25/27 Implications o First, in the digital economy era, companies are adjusting prices more frequently, and the effectiveness of traditional monetary policies has declined. Central banks need to pay attention to the influence of online markets on the monetary policy transmission, accelerate monetary policy transformation, and attach importance to high-frequency online inflation indicators. o Second, central banks should pay attention to the development and application of structural monetary policy in the future, and be more cautious in the application of aggregate monetary policy. Besides, it should also try to stagger the monetary policy adjustment cycle and price adjustment cycle appropriately, and continue to innovate monetary policy tools.
2020/12/29 26/27 Thank you!
Your questions are welcome!
Email: [email protected]
2020/12/29 27/27