Working papers 2020 Back to index

No posts.

  • Strategic Ambiguity in Global Games


    In incomplete information games with ambiguous information, rational behavior depends on fundamental ambiguity (ambiguity about states) and strategic ambiguity (ambiguity about others’ actions). We study the impact of strategic ambiguity in global games, which is evident when one of the actions yields a constant payoff. Ambiguous-quality information makes more players choose this action, whereas (unambiguous) low-quality information makes more players choose an ex-ante best response to the uniform belief over the opponents’ actions. If the ex-ante best-response action yields a constant payoff, sufficiently ambiguous-quality information makes most players choose this action, thus inducing a unique equilibrium, whereas sufficiently low-quality information generates multiple equilibria. In applications to financial crises, we demonstrate that news of more ambiguous quality triggers a debt rollover crisis, whereas news of less ambiguous quality triggers a currency crisis. 



    Consider an incomplete information game with players who have ambiguous beliefs about a payoff-relevant state. Players receive signals about a state, but they do not exactly know the true joint distribution of signals and a state. In this game, players’ beliefs about the opponents’ actions are also ambiguous even if players know the opponents’ strategies, which assign an action to each signal, because their beliefs about the opponents’ signals are ambiguous. Thus, rational behavior depends not only on fundamental ambiguity (ambiguity about states) but also on strategic ambiguity (ambiguity about others’ actions).



  • Decentralizability of Efficient Allocations with Heterogenous Forecasts


    Do price forecasts of rational economic agents need to coincide in perfectly competitive complete markets in order for markets to allocate resources efficiently? To address this question, we define an efficient temporary equilibrium (ETE) within the framework of a two period economy. Although an ETE allocation is intertemporally efficient and is obtained by perfect competition, it can arise without the agents forecasts being coordinated on a perfect foresight price. We show that there is a one dimensional set of such Pareto efficient allocations for generic endowments.  



    Intertemporal trade in complete markets is known to achieve Pareto efficiency when the price forecasts of agents coincide and are correct. The usual justification for this coincidence of price forecasts is that if agents understand the market environment perfectly,  they ought to reach the same conclusions, and hence in particular, their price forecasts must coincide. But it is against the spirit of perfect competition to require that agents should understand the market environment beyond the market prices they commonly observe; we therefore study intertemporal trade without requiring that price forecasts of heterogenous agents coincide.  



  • Imperfect Information, Heterogeneous Demand Shocks, and Inflation Dynamics


    Using sector-level survey data for the universe of Japanese firms, we establish the positive co-movement in the firm’s expectations about aggregate and sector-specific demand shocks. We show that a simple model with imperfect information on the current aggregate and sector-specific components of demand explains the positive co-movement of expectations in the data. The model predicts that an increase in the relative volatility of sector-specific demand shocks compared to aggregate demand shocks reduces the sensitivity of inflation to changes in aggregate demand. We test and corroborate the theoretical prediction on Japanese data and find that the observed decrease in the relative volatility of sector-specific demand has played a significant role for the decline in the sensitivity of inflation to movements in aggregate demand from mid-1980s to mid-2000s.



    A large class of macroeconomic models builds on the premise that firms set prices to fulfil demand. Several studies show that shocks to demand are heterogeneous and reflect aggregate and sector-specific disturbances.1 Knowing the source that originates the change in demand is important for setting the price consistent with profit maximization. Ball and Mankiw (1995) shows that the price should adjust if the change in demand originates from the aggregate shock, but it should remain unchanged if the change originates from the sector-specific shock. In reality, firms cannot observe the source of any change in demand. They therefore form expectations of aggregate and sector-specific components of demand based on the observed total demand and accrued knowledge from past aggregate and sector-specific shocks. Our analysis establishes important empirical regularities about firms’ expectations on the different aggregate and sector-specific components of total demand, it develops a parsimonious model of imperfect information that explains the co-movements in the expectations, and it studies the implications of empirically-congruous expectations for the sensitivity of inflation to changes in aggregate demand.



  • Japan’s Voluntary Lockdown: Further Evidence Based on Age-Specific Mobile Location Data


    Changes in people's behavior during the COVID-19 pandemic can be regarded as the result of two types of effects: the "intervention effect" (changes resulting from government orders or requests for people to change their behavior) and the "information effect" (voluntary changes in people's behavior based on information about the pandemic). Using mobile location data to construct a stay-at-home measure for different age groups, we examine how the intervention and information effects differ across age groups. Our main findings are as follows. First, the age profile of the intervention effect of the state of emergency declaration in April and May 2020 shows that the degree to which people refrained from going out was smaller for older age groups, who are at a higher risk of serious illness and death, than for younger age groups. Second, the age profile of the information effect shows that, unlike the intervention effect, the degree to which people stayed at home tended to increase with age for weekends and holidays. Thus, while Acemoglu et al. (2020) proposed targeted lockdowns requiring stricter lockdown policies for the oldest group in order to protect those at a high risk of serious illness and death, our findings suggest that Japan's government intervention had a very different effect in that it primarily reduced outings by the young, and what led to the quarantining of older groups at higher risk instead was people's voluntary response to information about the pandemic. Third, the information effect has been on a downward trend since the summer of 2020. While this trend applies to all age groups, it is relatively more pronounced among the young, so that the age profile of the information effect remains upward sloping, suggesting that people's response to information about the pandemic is commensurate with their risk of serious illness and death.



    The number of COVID-19 infections in Japan began to increase in earnest in the latter half of February, and by the end of March, the cumulative number of infections had reached 2,234. In response to the spread of infections, the government declared a state of emergency on April 7 for seven prefectures including Tokyo, and on April 16, the state of emergency was expanded to cover all prefectures. As a result, people refrained from going out, and the number of new infections in Japan, after peaking at 720 on April 11, began to drop, falling to almost zero by the end of May. This was the first wave of infections. However, in July, the number of new infections began to increase again, and continued to increase throughout the summer (peaking at 1,605 new infections on August 7). This was the second wave. While the second wave had subsided by the end of August, the number of new infections began to increase once again in late October, and on December 31, 2020, the number of new infections in Tokyo reached 1,353, exceeding 1,000 for the first time (the number of new infections nationwide was 4,534). In response, the government again declared a state of emergency on January 7. We are currently in the middle of the third wave.



  • The Welfare Implications of Massive Money Injection: The Japanese Experience from 2013 to 2020


    This paper derives a money demand function that explicitly takes the costs of storing money into account. This function is then used to examine the consequences of the large-scale money injection conducted by the Bank of Japan since April 2013. The main findings are as follows. First, the opportunity cost of holding money calculated using 1-year government bond yields has been negative since the fourth quarter of 2014, and most recently (2020:Q2) was -0.2%. Second, the marginal cost of storing money, which was 0.3% in the most recent quarter, exceeds the marginal utility of money, which was 0.1%. Third, the optimum quantity of money, measured by the ratio of M1 to nominal GDP, is 1.2. In contrast, the actual money-income ratio in the most recent quarter was 1.8. The welfare loss relative to the maximum welfare obtained under the optimum quantity of money in the most recent quarter was 0.2% of nominal GDP. The findings imply that the Bank of Japan needs to reduce M1 by more than 30%, for example through measures that impose a penalty on holding money.


    Seven years have passed since the Bank of Japan (BOJ) welcomed Haruhiko Kuroda as its new Governor and started a new regime of monetary easing, which was nicknamed the “Kuroda bazooka.” The policy goal that the BOJ set itself was to overcome deflation. At the time, the year-on-year rate of change in Japan’s consumer price index (CPI) was -0.9% and had been below zero for a long time. The measure the BOJ chose to escape deflation was to print lots of money. That is, the BOJ thought that it would be possible to overcome deflation by increasing the quantity of money. Specifically, in April 2013, the BOJ announced that it would double the monetary base within two years and thereby raise the CPI inflation rate to 2%. However, currently, CPI inflation remains stuck at 0.3%. The BOJ has not achieved its target of 2%, and there is little prospect that it will be achieved in the near future. While it is true that inflation currently is heavily affected by the sharp fall in aggregate demand since the outbreak of the COVID crisis in February 2020, which is putting downward pressure on prices, even before the crisis CPI inflation was only between 0.2 and 0.8% and therefore below the BOJ’s target.




  • Japan’s Voluntary Lockdown


    Japan’s government has taken a number of measures, including declaring a state of emergency, to combat the spread COVID-19. We examine the mechanisms through which the government’s policies have led to changes in people’s behavior. Using smartphone location data, we construct a daily prefecture-level stay-at-home measure to identify the following two effects: (1) the effect that citizens refrained from going out in line with the government’s request, and (2) the effect that government announcements reinforced awareness with regard to the seriousness of the pandemic and people voluntarily refrained from going out. Our main findings are as follows. First, the declaration of the state of emergency reduced the number of people leaving their homes by 8.6% through the first channel, which is of the same order of magnitude as the estimate by Goolsbee and Syverson (2020) for lockdowns in the United States. Second, a 1% increase in new infections in a prefecture reduces people’s outings in that prefecture by 0.026%. Third, the government’s requests are responsible for about one quarter of the decrease in outings in Tokyo, while the remaining three quarters are the result of citizens obtaining new information through government announcements and the daily release of the number of infections. Our results suggest that what is necessary to contain the spread of COVID-19 is not strong, legally binding measures but the provision of appropriate information that encourages people to change their behavior.


    In response to the spread of COVID-19, the Japanese government on February 27 issued a request to local governments such as prefectural governments to close schools. Subsequently, the Japanese government declared a state of emergency on April 7 for seven prefectures, including Tokyo, and on April 16 expanded the state of emergency to all 47 prefectures. Prime Minister Abe called on citizens to reduce social interaction by at least 70% and, if possible, by 80% by refraining from going out. In response to these government requests, people restrained from going out. For example, in March, the share of people in Tokyo leaving their homes was down by 18% compared to January before the spread of COVID-19, and by April 26, during the state of emergency, the share had dropped as much as 64%. As a result of people refraining from leaving their homes, the number of daily new infections in Tokyo fell from 209 at the peak to two on May 23, and the state of emergency was lifted on May 25.




  • Consumer Inventory and the Cost of Living Index: Theory and Some Evidence from Japan


    This paper examines the implications of consumer inventory for cost-of-living indices (COLIs) and business cycles. We begin by providing stylized facts about consumer inventory using scanner data. We then construct a quasi-dynamic model to describe consumers’ purchase, consumption, and inventory behavior. A key feature of our model is that inventory is held by household producers, not by consumers, which enables us to construct a COLI in a static manner even in an economy with storable goods. Based on this model, we show that stockpiling during temporary sales generates a substantial bias, or so-called chain drift, in conventional price indices, which are constructed without paying attention to consumer inventory. However, the chain drift is greatly mitigated in our COLI, which is based on consumption prices (rather than purchase prices) and quantities consumed (rather than quantities purchased). We provide empirical evidence supporting these theoretical predictions. We also show empirically that consumers’ inventory behavior tends to depend on labor market conditions and the interest rate.


    Storable goods are abundant in the real world (e.g., pasta, toilet rolls, shampoos, and even vegetables and milk), although most economic models deal with perishable goods for the sake of simplicity. Goods storability implies that purchases (which are often observable) do not necessarily equal consumption (which is often unobservable), and the difference between the two serves as consumer inventory. In particular, temporary sales and the anticipation of an increase in the value-added tax rate often lead to a greater increase in purchases than consumption. Moreover, the COVID-19 outbreak in 2020 caused many products, such as pasta and toilet rolls, to disappear from supermarket shelves, which would not have happened if these products were not storable. The stockpiling behavior by consumers poses challenges for economists, for example in the construction of price indices. 





  • Efficiency, Quality of Forecasts and Radner Equilibria


    We study a simple two period economy with no uncertainty and complete markets where agents trade based on forecasts about the second period spot price. We propose as our solution concept a set of forecasts with the following properties: there exist (heterogenous) forecasts contained in this set that lead to efficient allocations, the set contains only those forecasts that correspond to some efficient equilibrium, and finally that the forecasts assign positive probability to the actual market clearing spot price. We call such a set of prices an efficient equilibrium with ambiguity, and interpret it as a generalization of Radner equilibrium that delivers efficient allocations under forecasts that possess a self-fulfilling property that is weaker than perfect foresight. 


    Walrasian trade in intertemporal economies require households to forecast prices that will prevail in spot markets at future dates. The ubiquitous nancial equilibrium model that is used to address this aspect of intertemporal economies is the one proposed by Radner (1972) (following Arrow (1963)) and is the bedrock of modern treatments of general equilibrium. This resulting Radner equilibrium (henceforth, RE), postulates that households correctly anticipate all spot prices at future dates; a RE is accordingly an equilibrium with perfect foresight (henceforth, PFE), where the forecasts of heterogenous households are perfectly aligned.




  • Online Consumption During the COVID-19 Crisis: Evidence from Japan


    The spread of novel coronavirus (COVID-19) infections has led to substantial changes in consumption patterns. While demand for services that involve face-to-face contact has decreased sharply, online consumption of goods and services, such as through e-commerce, is increasing. The aim of this study is to investigate whether online consumption will continue to increase even after COVID-19 subsides, using credit card transaction data. Online consumption requires upfront costs, which have been regarded as one of the factors inhibiting the diffusion of online consumption. However, if many consumers made such upfront investments due to the coronavirus pandemic, they would have no reason to return to offline consumption after the pandemic has ended, and high levels of online consumption should continue. Our main findings are as follows. First, the main group responsible for the increase in online consumption are consumers who were already familiar with online consumption before the pandemic and purchased goods and service both online and offline. These consumers increased the share of online spending in their spending overall and/or stopped offline consumption completely and switched to online consumption only. Second, some consumers that had never used the internet for purchases before started to use the internet for their consumption activities due to COVID-19. However, the share of consumers making this switch was not very different from the trend before the crisis. Third, by age group, the switch to online consumption was more pronounced among youngsters than seniors. These findings suggest that it is not the case that during the pandemic a large number of consumers made the upfront investment necessary to switch to online consumption, so a certain portion of the increase in online consumption is likely to fall away again as COVID-19 subsides. 


    People’s consumption patterns have changed substantially as a result of the spread of the novel coronavirus (COVID-19). One such change is a reduction of the consumption of services that involve face-to-face (F2F) contact. For instance, “JCB Consumption NOW” data, credit card transaction data provided jointly by JCB Co., Ltd. and Nowcast Inc., show that, since February this year, spending on eating out, entertainment, travel, and lodging have shown substantial decreases. Even in the case of goods consumption, there has been a tendency to avoid face-to-face contact such as at convenience stores and supermarkets. For example, with regard to supermarket shopping, the amount of spending per consumer has increased, but the number of shoppers has decreased. Another important change is the increase in the consumption of services and goods that do not involve face-to-face contact. The credit card transaction data indicate that with regard to service consumption, spending on movies and theaters has decreased substantially, while spending on content delivery has increased. As for the consumption of goods, so-called e-commerce, i.e., purchases via the internet, has shown substantial increases.  




  • コロナ収束後もオンライン消費の増加は続くか クレカ取引データを用いた分析


    新型コロナの感染拡大に伴い人々の消費スタイルが大きく変化している。外食や娯楽などFace-to-face の接触を伴うサービスへの需要が激減する一方,E コマースなどモノやサービスのオンライン消費は増えており,コロナ収束後も続くとの見方がある。ポストコロナはコロナ前に戻るのではなく,オンライン消費を軸に新たな消費スタイルが生まれるとの見方もある。


    本稿では以下のファインディングを得た。第1 に,オンライン消費増加の主たる担い手は,コロナ前からオンライン消費に馴染み,オンラインとオフラインを併用していた消費者である。こうした消費者が,オンライン消費の割合を高め,さらにはオフライン消費を一切やめてオンラインのみに切り替えた。第2 に,オンライン消費の経験のない消費者の一部が,コロナを機にオンライン消費を始める動きもみられた。ただし,その度合いはコロナ前のオンライン化の趨勢と大きく異ならなかった。第3 に,年齢別にみると,若年層がオンライン消費を増やした一方,シニア層の寄与は小さかった。例えば,コロナ前にオンラインとオフラインを併用していた20 代後半の消費者のうち16%がオンライン消費のみに切り替えたが,同じくコロナ前にオンラインとオフラインを併用していた60 代前半の消費者のうちでオンラインのみに切り替えたのは11%であった。オンライン消費への切り替えの年代間の差は,デジタルリタラシーの差によるものではなく,感染を回避する姿勢の差を反映していると考えられる。


    1 ポストコロナの個人消費

    新型コロナの感染に伴い人々の消費スタイルが大きく変化している。ひとつはFace-to-face の接触を伴うサービスの消費を抑える動きである。『JCB 消費NOW』でも本年2 月以降,外食や娯楽,旅行,宿泊が大幅な減少を示している。モノ消費でも,コンビニやスーパーの店頭でのFace-to-face の接触を嫌う傾向がある。例えば,スーパーでの購買は,1 人の消費者が購買する金額は増えており,そのためスーパーでの購買金額は増えているものの,購買者数自体は減っている。
    もうひとつの重要な変化は,Face-to-face の接触を伴わないサービスやモノの消費の拡大である。『JCB 消費NOW』でみると,サービス消費では,映画や劇場での消費が大幅に減少する一方,コンテンツ配信は増えている。モノ消費についても,ネット経由での購買,いわゆるE コマースが大幅な伸びを示している。




  • How Much Did People Refrain from Service Consumption due to the Outbreak of COVID-19?


    With the spread of coronavirus infections, there has been a growing tendency to refrain from consuming services such as eating out that involve contact with people. Self-restraint in service consumption is essential to stop the spread of infections, and the national government as well as local governments such as the Tokyo government are calling for consumers as well as firms providing such services to exercise self-restraint. One way to measure the degree of self-restraint has been to look at changes in the flow of people using smart phone location data. As a more direct approach, this note uses credit card transaction data on service spending to examine the degree to which people exercise self-restraint. The results indicate that of men aged 35-39 living in the Tokyo metropolitan area, the share that used their credit card to pay for eating out in March 2020 was 27 percent. Using transaction data for January, i.e., before the full outbreak of the virus in Japan, yields an estimated share of 32 percent for March. This means that the number of people eating out fell by 15 percent. Apart from eating out, similar self-restraint effects can be observed in various other sectors such as entertainment, travel, and accommodation. Looking at the degree of self-restraint by age shows that the self-restraint effect was relatively large among those in their late 30s to early 50s. However, below that age bracket, the younger the age group, the smaller was the self-restraint effect. Moreover, the self-restraint effect was also small among those aged 55 and above. Further, the degree of self-restraint varies depending on the type of service; it is highest with regard to entertainment, travel, and accommodation. The number of people who spent on these services in March 2020 was about half of the number during normal times. However, the 80 percent reduction demanded by the government has not been achieved.


    With the spread of coronavirus infections, there has been a growing tendency to refrain from consuming services such as eating out that involve contact with people. Self-restraint in service consumption is essential to stop the spread of infections, and the national government as well as local governments such as the Tokyo government are calling for consumers as well as firms providing such services to exercise self-restraint.
    Specifically, Prime Minister Shinzo Abe declared a one-month long state of emergency in Tokyo and six other prefectures on April 7, 2020 and expanded it to the entire country on April 16. PM Abe stated in his speech on April 7 that “According to an estimate by the experts, if all of us make efforts and reduce opportunities for person-to-person contact by a minimum of 70 percent, or ideally 80 percent, we will cause the increase in the number of patients to reach its peak two weeks from now and shift over into a decrease. . . . I ask people to refrain from going out, aiming at a 70 to 80 percent decrease, for the limited period of one month between now and the end of Golden Week holidays on May 6.”
    The purpose of this note is to measure the degree to which people in Japan have been exercising self-restraint since the outbreak of COVID-19. One way to do so is to look at changes in the flow of people using mobile phone location data.2 As a more direct approach, we use credit card transaction data on service spending to examine the degree to which people exercise self-restraint.