無形投資和低通貨膨脹-一個框架和一些證據(jù)

1、ContentsAbstract 2Introduction 4A Few Important Macroeconomic Puzzles 5Can the AS-AD Framework Still Explain Low Inflation? 6Empirical Support for More Elastic Aggregate Supply 9Can Intangible Investment Explain the AS Curve flattening? 14What Have Been Driving the Flattening of the AS Curve? 16Conc

2、lusions 20References 22Appendix 25INTRODUCTIONFollowing the Global Financial Crisis (GFC) of 2007-08, there has been a lively debate among economists about a number of prominent features of the subsequent recovery. An important question for macroeconomists and policymakers has been why despite the m

3、assive macroeconomic stimulus that was injected at the time, many economies have struggled to achieve their inflation goals even as full employment was eventually restored. While monetary policy remained highly accommodative across most advanced economies, inflation remained largely below (or near)

4、the target despite full employment and closed output gaps.2 A related question is why despite historically low unemployment, wage pressures did not emerge in any significant way. This has sometimes been linked to the declining bargaining power of labor, the increasing concentration and monopolistic

5、positions of firms in both “tech” and “non-tech” sectors, and structural changes that have led to redistribution of income from capital to labor.In this paper, we present a simple framework to argue that the rise of intangible investment across advancedand increasingly emergingeconomies can plausibl

6、y explain many macroeconomic relationships observed over the past decade. For purposes of this analysis, intangible investment refers to the nonphysical investment in the production of economic output. While software is a prominent and well-known example of intangible investment, other examples such

7、 as branding, production and managerial processes, and training also have a meaningful impact on the functioning of the economy. This type of investment is becoming a rising share of total investment and in some economies already exceeds the share of tangible investment.We focus on the properties of

8、 intangible investment that are different from conventional tangible investment (for example, “plants and equipment” or hardware) and that can modify the conventional understanding of macroeconomic equilibria and shocks. We argue that the underlying structural changes that intangible investments emb

9、ody, while in train since at least the 1990s, are still at an early stage of transforming economies. Despite that these changes are still at an early stage both within and across countries, we present both time series and panel data evidence that the rise of intangible investment is already having a

10、 measurable impact on macroeconomic relationships. In particular, this papers conceptual and empirical focus is on the changing relationship between output, inflation and unemployment. The paper also discusses the implications of the conceptual framework on understanding inequality and wage stagnati

11、on, the rise of monopolistic power, financing investment, and taxation of investment, which is the focus of parallel work.This paper is structured as follows. The next section describes a few economic phenomena that have puzzled economists and policymakers over the past decade and the related litera

12、ture.2 The analysis in this paper refers to the period prior to the onset of the Covid-19 pandemic of 2020.Section III uses the simple Aggregate Supply-Aggregate Demand (AS-AD) framework to illustrate how persistently low inflation would be represented in such a framework and argues that a more elas

13、tic aggregate supply curve than in the past would be conceptually consistent with the empirical observations. Section IV provides more rigorous empirical support to the proposition that aggregate supply curves may be more elastic than is traditionally assumed. The subsequent section discusses the ri

14、se of intangible investment and the unique properties that differentiate it from tangible investment. Section VI provides empirical evidence supporting the papers main thesis that intangible investment can help explain low inflation. The final section discusses some addition implications of the prop

15、osed framework and topics of parallel research.A FEW IMPORTANT MACROECONOMIC PUZZLESA major economic question since the global financial crisis has been about the causes underlying low inflation. Even as output gaps closed since the crisis and unemployment recovered to record low levels prior to the

16、 onset of the Covid-19 related downturn in 2020, inflation in some cases remained stubbornly below the pre-crisis norm of advanced economies and central banks target. More recently, emerging markets also began to experience low inflation.3 Explanations regarding low inflation have focused on demogra

17、phic changes, technological changes, globalization, the expansion of central bank independence and inflation targeting, as well as Fisherian low interest rate-low inflation relationships.4 A related strand of the literature has tried to explain the breakdown of the relationship between unemployment

18、and inflation since the 1990s.5 This literature has not achieved consensus on the causes for the so-called flattening of the Philipps curve, and consequently little guidance for the likely relationship in the future.In terms of explanations for why technological progress in particular leads to lower

19、 inflation, numerous explanations have been offered. Then Federal Reserve Chairman Alan Greenspan in congressional testimony suggested that technological progress has suppressed unit labor costs and increased productivity (2005). In general, the discussion of technology has centered around three asp

20、ects: the price of information and communication technology (ICT) inputs has declined, leading to a decline of overall inflation due to the weight of ICT in the index; technology has had an impact through competition and market structure due to the ability to compare prices worldwide instantaneously

21、 and transparently; and technological change suppresses wage3 See Ha, Kose and Ohnsorge (2019) for a comprehensive review of the experience of emerging and developing economies.4 Sanchez and Kim (2018) discuss several hypotheses put forward in the literature and policy discussions including notably

22、that technological progress may have helped lower inflation by suppressing unit labor costs and elevating productivity growth. Other explanations they discuss include demographic transition, globalization, the expansion of central bank independence and inflation targeting, and a Fisherian low intere

23、st rates-low inflation nexus.5 See Gagnon and Collins (2019) for a discussion and a review of the main hypotheses for why the Phillips curve relationship may appear to have broken down.growth.6 One shortcoming with these explanations is that no single good theoretical framework has emerged on why te

24、chnological progress has had such a big impact on inflation, even as its impact on productivity and output growth remains elusive.The suppression of wage growth across many countries over past decades has been associated with high and rising inequality in both advanced and emerging economies. This h

25、as been increasingly linked to technological change in empirical studies.7 Other explanations for rising inequality included the weaker bargaining power of labor8, the rise of artificial intelligence and automation9. Moreover, explanations including weaker bargaining power of labor are highly plausi

26、ble for a range of countries, but the deeper underlying systematic driver of this reduced power remains less clear i.e. possibly due to country-specific legislation, ability of firms to shift production offshore, or the threat of substituting human labor with technology.In summary, while low inflati

27、on, an increasing disconnect between unemployment and inflation, and high and rising inequality are well established, incorporating the structural changes in economies that lead to this outcome in a simple stylized macroeconomic framework that would apply to a broad group of economies have proved el

28、usive.10 The next section attempts to illustrate how persistently low inflation might conceptually present itself in a macroeconomic framework that also allows for shocks.CAN THE AS-AD FRAMEWORK STILL EXPLAIN LOW INFLATION?Counter-cyclical macroeconomic policies can often be well interpreted under t

29、he simple and standard AS-AD framework (Figure 1). In this framework, the ideal equilibrium for the economy is point E*, where the vertical long-term supply curve LS, which indicates the potential output Y*, intercepts with both the downward sloping aggregate demand (AD) curve and the upward sloping

30、 short-term aggregate supply (AS) curve. A large negative aggregate demand shock6 Lv, Liu and Xu (2019) find evidence that technological progress and globalization have helped reduce inflation in the United States. Autor et al link the rise of superstar firms to wage suppression.7 Jaumotte, Lall and

31、 Papageorgiou (2008, 2013) find that globalization had much less of a role to play than earlier studies had suggested. Instead, technological change can exacerbate inequality by increasing the wages of those with complementary skills much more tangibly.8 Bental and Demougin (2010) demonstrate how th

32、e institutional design of labor market bargaining can reduce the labor share of national income. Stansbury and Summers (2020) explain sluggish wage growth, lower inflation and a declining share of labor income in the United States through reduced worker power which they estimate using a rent-sharing

33、 parameter between workers and shareholders.9 Leduc and Liu (2019) find evidence of automation contributing substantially to the decline in the share of labor income in the United States.10 Farhi and Gourio (2018) extend a neoclassical growth model and using US data identify a role for monopoly powe

34、r in explaining some of these trends in the United States.would knock the economy into recession (E), where output (Y) and price (P) both fall far below the potential or target levels. In this situation, counter-cyclical macro measures, including loose monetary policy and/or fiscal stimulus, would h

35、elp to lift the economy out of recession by boosting the aggregate demand. If the economy only goes partially back (E) towards the long- term equilibrium, it is often an indication of continued needs for accommodative macro policy support.Persistent low inflation, however, has been a puzzling incons

36、istency between the AS- AD model and many economies actual experience in recent years. Figure 1 suggests that, when the economy recovers from a severe downturn, inflation should rise back to its pre-crisis equilibrium level as growth approaches the potential.Nonetheless, what happened in many countr

37、ies since the GFC is that, while stimulus measures have helped to narrow or even close the negative output gap, inflation has remained stubbornly low.Some researchers worry that macro economics has entered an uncharted territory, with important discussions such as whether inflation targeting is stil

38、l anappropriate policy framework left unanchored.Figure 1. The AS-AD FrameworkPLSASP*E*EPEADPADADYYY*YIs the AS-AD framework really outdated? No. In fact, its discrepancy with the recent reality can be largely reconciled with a simple tweak: supply side changes. In Figure 2, in addition to (a) the i

39、nitial shock that shifted the aggregate demand from AD to AD and (b) the subsequent policy stimulus that pushed it back to AD, it is also assumed that the aggregate supply curve has flattened in the meanwhile, from AS to AS. With this modification where the economys short term supply function is mor

40、e elastic, price level P would remain well below its pre-crisis equilibrium P*, even when the output gap is fully closedY is the same as Y*.Introducing supply side changes not only helps the AS-AD model better fit reality, but also brings profound policy implications. While the dynamics illustrated

41、in Figure 1 cannot square persistent low inflation with a closed output gap, adding supply side changes solves the puzzle, at least theoretically. The natural questions to ask next would be: (a) whether the short-term AS curve has indeed become more elastic in recent years; and (b) if so, what the u

42、nderlying driving factors have been.Before empirically addressing these questions, however, it is worth first highlight some key policy implications of the adjusted AS-AD analysis:Figure 2. The AS-AD Framework with Flattened AS CurvePLSASP* EASPEEADPADADYY* (Y)YLow inflation could be a new long-term

43、 equilibrium. Note that in Figure 2, E is not only the intersection of the AD curve AD and the short-term AS curve AS, but also their intersection with the long-term aggregate supply curve LS. This suggests that, rather than indicating slack or imbalance in the economy, the lower price level P is a

44、new long-term equilibrium.A path of stagnation to inflation target? The analysis does not directly answer the question whether the inflation target, set at P*, is still a proper policy anchor, but it does shed more light on likely paths through which the economy could hit the inflation target.The pa

45、th illustrated in Figure 3 has two phases, the expansion indicated by the red arrows and the stagnation indicated by the blue arrows. At E, since inflation is still below the target level, committed policy makers will continue with supportive measures, which not only pushes growth beyond the potenti

46、al but also leads to higher inflation (expansion). If the policy makers have perfect foresight, policy stance should turn neutral once the aggregate demand curve reaches AD. What happens next is that, as growth stays above the potential and as near-term price rigidity diminishes over time, inflation

47、 will gradually rise to the target level P* while the aggregate supply curve shifts from AS to AS. At this stage following the expansion, inflation will keep rising but growth will slow (stagnation).This contrasts Figure 1 where the economy from E would reach the inflation target through continued e

48、xpansion rising inflation accompanied by faster growth.Achieve the inflation target yet avoid stagnation? There are indeed such possibilities, for instance, the path shown in Figure 4. If potential growth rises to Y#, either due to the flattening of the AS curve or not, continued policy support coul

49、d eventually move the AD curve from AD to AD#. In this case, the new long-term supply curve LS#, the AD curve AD# and the short-term AS curve AS will together settle the economy at the new long-term equilibrium E#, with target level of inflation P* but higher growth Y# than before.The different like

50、ly paths to achieving inflation target highlights a critical judgement for policy makers, that is, whether and how potential growth has changed over time. We will defer further discussions on this, including the consequence of misjudgment, to the final section where we revisit policy issues while al

51、so considering the empirical findings presented in the next two parts.Figure 3. Likely Path to Inflation Target:StagnationFigure 4. Likely Path to Inflation Target:Higher Potential GrowthPLSASASP* E*ASPEADADPLSLS#P*E*E#AD#PEADASADASY* (Y)YY* (Y)Y#YThe initial aggregate demand shock was omitted for c

52、larity.The initial aggregate demand shock was omitted for clarity.EMPIRICAL SUPPORT FOR MORE ELASTIC AGGREGATE SUPPLYThe previous section demonstrates that by introducing a “flatter” AS curve, the AS-AD framework resolves a major puzzle observed in recent years, that is, persistent low inflation alo

53、ng with a narrowed or even closed output gap. Such an adjustment to the model is more than a theoretical tweak but, as shown below, captures actual macro development in many economies.There is a large literature documenting the flattening of the Phillips curve, which is closely associated with a fla

54、ttened AS curve in the AS-AD framework.11 An influential study is IMF (2013). It finds that the slope of the Phillips curve has gradually flattened over the past several decades across a large sample of advanced economies. An earlier study, IMF (2006), is another cross-country study showing that sen

55、sitivity of prices to domestic economic cycle has declined over time. It suggests globalization as an important a driver underlying such developments.There have also been studies looking at individual economies, for instance, Blanchard (2016) and Ball and Mazumder (2010) focus on the US economy, whi

56、le Iakova (2007) discusses the likely implications for UK monetary policies.12It should be pointed out that the discussions surrounding the flattening of the Phillips curve have not been unequivocally settled. There are still counterarguments, often resting on the endogeneity among economic developm

57、ent, market expectations and macroeconomic policies, arguing that the Phillips curve has not flattened. A notable paper is McLeay and Tenreyro (2019). Aside from the degree of the AS curve flattening, no consensus has been reached either regarding the timing of the flattening and what might have bee

58、n the underlying drivers.Rather than attempting to settle this debate, the evidence below is intended to highlight the empirical relevance of a flattened AS curve given its policy significance.A visual illustrationFigure 5 plots the inflation gap in the US and Germany against their respective output

59、 gap.13 In both economies, the AS curve for the 1990s and early 2000s (the red dash line) was clearly steeper than the one for the latter period of the sample (the blue line). Similar charts for a larger set of countries are presented in Figure 6. While not universally the case (for instance, Japan

60、and Mexico are two exceptions), a flattened AS curve to varying degrees is indeed observed in most economies.11 The conventional Phillips curve describes the negative relationship between wage inflation and unemployment. Assuming a stable Okun link between output growth and unemployment, the flatten