周末荐读 | 前IMF驻华代表李一衡：全球增长放缓与负实际利率

观点速递

本文作者是IMI学术委员、韩国对外经济政策研究院（KIEP）院长李一衡。本文关注目前部分发达经济体正面临负利率环境的原因及这种“非正常”情况被后的经济特性。

文章首先辨明负利率的概念，从四个方面讨论了负利率大行其道的原因。接着，作者提出了当前实际利率与自然实际利率，作者分析了自然实际利率走低的两个原因。本文认为伴随着过去措施储蓄机会的老龄化、债务积压及流动性过剩是压低自然利率的主要原因。由于全球化、技术变革和金融化的促进和宽松的政策的巩固，世界经济曾呈现强势复苏态。前两种原因虽然算不上造成低增长问题的原因，却上放缓全球经济复苏的始作俑者。

最后作者提出应对目前低增长陷阱有两个办法：压低实际利率，或保持利率而让过去失去养老储蓄机会的家庭采取行动弥补。而目前无从得知哪种方法成效更好。但是如果目前的问题本质上是结构性问题，前种方法只能推迟必要的改革，并引发更大的问题。在两种情况下，实际利率都很有可能维持低于历史平均的水平。但是，将实际利率通过利率政策降到了极大的负值区间，可能会导致自然利率降到更低的风险。

中文译文如下：

全球增长放缓与负实际利率

李一衡

翻译：喻凡

根据一份十国集团（下称G10）的研究[1]，负实际利率只在两个时期长期存在过，大致在两次大战时期，即20世纪20年代和40年代。这份研究提出两个问题：为什么大部分发达经济体正面临着负利率环境，以及这种明显的“非正常”情况背后存在何种经济特性。本文认为，伴随着过去错失储蓄机会的老龄化、债务积压及流动性过剩是压低自然利率的主要原因。由于全球化、技术变革和金融化的促进和宽松货币政策的巩固，世界经济复苏曾势头强劲。前两种因素虽算不上造成低增长问题的原因，却是放缓全球经济复苏的始作俑者。

“

负实际利率的含义

利率是持有货币的机会成本，如，可以通过一种无风险债券的收益率来衡量。由于持有一种金融资产意味着将支出（现金）延迟至未来某一天，而供其他人当前支出使用或者投资，其又可被定义为代内或代际资源转移。（由非金融机构持有的）金融资产可能因审视的角度不同而有不同含义；从资源转移视角[1]看，它是一种实现转移的重要工具。个人之间的代内转移可以描述为将一名持有人的资产出借给另外一个人，而社会中的代际转移则是出借资产给政府，此时所有的资产持有人都承有还款负担。

因为利率是持有金融资产的收益，均衡状态下，利率也相当于资本(收益)成本。从公平角度看，西尔沃·格塞尔（Silvio Gesell）[2]主张将所有金融资产设置零利率，因为资产的持有者不应比那些被迫借债（没有资产）的人享有更低成本资本优势[3]。然而，本文没有进入这个方向的讨论，以下是负实际利率可能大行其道的几种情况。

1)    如果经济环境极其不稳定以至未来收入难以保证，此时风险调整资本收益为负，则负实际利率可能会盛行。这种情况下，投资者不会自己持有资产，而会付钱给某人（政府）让他人使用其资产，以便至少能保留资产的原有价值，或者相较本人持有减少价值损失。[4]

2)    同样，如果让别人使用资产的情况盛行，当前所持有资产的未来价值会多于现值，即负贴现率。对于一个试图扩大终生消费的资产持有人，若其资产的净现值（未来消费）上升，那么他会储蓄得更少。

3)    一脉相承的是，如果劳动力规模预期缩减非常严重，甚至超过任何实质性的生产力增长，负利率也可能盛行。然而，在开放的经济体中，这有可能导致资本流动的再调整，即资本会流入预计会有劳动力（或生产力）增长的区域。

4)    从市场的观点看，如果一个经济体的债务存量大于其未来收入总量的净现值，该经济体会通过通缩去除债务——这是另一种增加税收的方法。[5]随着通胀的增长，债务价值的减少可能更有即时性，但是没有通胀时，负实际利率可充当去除债务存量的替代性方案（除非资产由于动荡而贬值）。

“

当前实际利率vs.自然实际利率

当前，发达经济体的长期实际利率虽不稳定，但总体呈下降趋势（图1）。现在与过去不同的是，低或负实际利率的产生是由于名义利率的下降，而不是高通胀的骤升。尽管实际利率应为零还是负尚无定论，但相比历史均值，其值确有降低。实际上，19世纪80年代至20世纪90年代期间，长期实际利率（即G10国家的十年期国债平减CPI后的加权平均值）[1]均值为1.5-2%，除去两次大战期间则为2.5-3%（图2）。两次大战期间的长期实际利率为何为负虽不是本文讨论的重点，但其原因可能包括资本收益为负，或人们愿意持有国债（一个资产保值最保险的办法）。

如G10研究准确指出的那样，发达国家的经济历史显示，正实际利率才是“常态”或市场决定的自然利率。更具体一点，以美国为例，美国的战后平均实际利率（十年短期国债）约为2.5，基本上相当于G10国家的平均值。虽然1974-1975，1979-1980及2011-2012期间，十年期政府债券收益率短暂为负，但大都是突发的通胀造成的。

“

为什么自然实际利率可能走低？

a. 过去错失的储蓄机会

发达经济体的储蓄占国内生产总值的比重在20世纪80年代末期达到峰值24%，之后持续下跌，在全球金融危机后大幅下跌至20%（图3）。这与过去50年内，全球平均储蓄与按购买力平价计算的国内生产总值的比例大致相同，这一比值约为22%。1世纪2开始该比重逐渐下降，同时也与预期储蓄行为和全球人口情况相符。图4展示了在欧洲和美国，如果家庭追求长期内的消费平稳化，储蓄率偏离平均水平的变化（基于联合国人口预测）。虽然预测的准确度值得商榷，但至少它大致描绘了在家庭理性的情况下储蓄的长期演变。也就是说，欧洲在20世纪80年代的储蓄本应达到最高水平，并大约从2020年开始降低。美国的情况类似，但有一点不同，即美国储蓄本应在21世纪前十年达到峰值。

过去几十年内，中国、日本、韩国的政府债券实际收益率[1]为正，并呈下降趋势（图5）。负实际利率出现在20世纪70年代的日本和韩国，一部分原因是油价和全球金融状况引发了高通胀。就中国而言，在20世纪80年代晚期和90年代中期，压低实际利率至负值的高通胀就是中国政府努力实现快速过渡的副产品。在最近几年中，日本和其他几个发达经济体一样，实际利率降为负值。2.5%的长期（1980-2016）平均数与美国该数字大致相似。韩国的实际利率更高，为5%，这反映了该时期韩国较强的经济发展态势。就中国而言，由于前面提到的过渡期，所以难以简单对比中国与其他国家的长期发展。但中国2000年后的平均实际利率约为3.5%（韩国为2%），这说明其资本收益率较高。

中日韩的实际储蓄率（图6）基本符合基于人口的储蓄预测(获取方式与上文发达经济体相同，图7是中日韩的储蓄预测情况)。由图7可知，中国的储蓄在2010年左右达到峰值，日本是1990左右，韩国则在1990-2000年间。与美国和欧洲不同，实际存款大致与预测相符。

然而，即使过去这几个国家没有错过存款的机会，也有两个因素可以推动中国和韩国（稍微推动日本）的储蓄。第一，相比一二十年前人们寿命的极大延长；第二，没有明显恢复趋势且更加不容乐观的经济形势。中韩经济很大程度依赖出口，所以全球经济萧条也意味着国内经济发展放缓。以上两个因素都是增加储蓄率的强劲动力，把中韩两国推到与美国、欧盟相似的境遇。中国的情况在图8中有所展示。

b. 债务积压和流动性过剩

图8的“其他债务”包括金融负债或公共债务（或两者之和，虽然本质上两者不能相加）。金融负债是代内债务，并且在运转情况良好的体系下，一个群体向另一个相同群体举债并不会产生问题。只有负债群体收入受限且偿还债务能力有限时才会造成风险。公共债务对消费有类似的影响，只不过这里的偿债负担更广地分散到未来的税收负担中。在过去二十年里，金融资产负债和政府债务占国内生产总值比重急剧上升（见图9、图10）。

金融资产和负债是一种现阶段持有某群体的储蓄存款、未来再进行清偿的方式。如果与生产力（即预期收入流）相比，金融资产过多，那极有可能导致存款价值受损。使存款价值受损的方式可能是资产泡沫的突然破裂、负实际利率的逐步影响、持续的高膨胀或是政治程序。如果金融资产成为一种系统性风险，风险会转移到生产力上，就像美国和欧盟在全球金融危机之后经历的那样。

“

结论：问题在于把低消费放在现在还是未来

如果说20世纪80年代到21世纪前20年，基于人口的消费平滑路径有任何指导意义的话，那就是提醒面临老龄化压力的家庭需要赶上来，也就是说，他们要储蓄高于长期平均水平的存款。但是这可能意味着抑制消费至长期平均水平之下，进而抑制投资。这反过来又会抑制经济发展，开启恶性循环。应对目前低增长陷阱有两个方法。

第一个方法是压低实际利率，使其低于当前较低的自然利率，甚至是达到负利率。这会导致资产持有人减少储蓄。然而，如果该陷阱的深层原因是本质上的结构性问题，这一方法只会产生短期效果，且经济状况会在几年之内恶化，这又需要进一步降低实际利率。另一方法是保持实际利率，而让过去失去养老储蓄机会的家庭采取行动弥补。这一办法在短期内会拖慢经济增长，但是一旦达到资本储蓄目标，消费就会恢复。

到底哪种方法成效更好，或者哪种能最小化终生消费的损失，都无从得知。但是，如果目前的问题本质上是结构性问题，前种方法只能推迟必要的改革，并引发更大的问题。如果我们放任市场，市场会自行修正，同时伴随着为更强劲的经济恢复打好基础的修正性结构改革，并使公共债务和金融资产会下降。在两种情况下，实际利率都很有可能维持低于历史平均的水平。但是，将实际利率通过利率政策降到极大的负值区间，可能会导致实际自然利率降到更低的风险。

英文原文如下：

Global Sluggish growth 

and negative real interest rate

Il Houng Lee

According to a G10 Study,[1] there were only two periods during which negative real interest rates prevailed for a prolonged period. These were during the 1920s and 1940s—broadly the two war periods. This observation raises the question as to why most of the advanced economies are currently facing real negative interest rates and what economic characteristics are behind this apparent “abnormal” situation. This note suggests aging coupled with foregone savings opportunity in the past, and debt and liquidity overhang to be the main causes of pushing down the neutral rate of interest. These are not so much the originators of the low growth problem, but rather factors that sloweddown the recovery of the global economy after a period of exuberant strong performance facilitated by globalization, technological changes and financialization and underpinned by accommodative policies.

“

Meaning of negative real interest rate[2]

Interest rate is the opportunity cost of holding money measured, for instant, by the yield of a risk free bond. Since holding a financial asset implies deferred spending (of cash) for a future datefor someone else to spend now or to invest, it can be defined as intra and inter-generational transfer of resources.Financial assets (held by non-financial institutions) may have different meaning depending on the perspective it is looked at, from a resource transfer perspective,[3] it is an important instrument for such transfer. Intra-generational transfer of individuals can be interpreted as lending one agent’s wealth to another while inter-generational transfer of a society as lending to the government where all agents take on the burden of repayment.

Since interest rate is the return on holding financial assets, at equilibrium, it is also equivalent to the cost of (return on) capital. From an equity point of view, Silvio Gesell[4] argued for zero interest rate on all financial assets since a holder of wealth should not be given the advantage of lower cost of capital relative to those who has to borrow (with no wealth).[5]Nevertheless, without going into this line of argument, below are possible cases where negative real interest rate could prevail.

i)   Negative real interest rate could prevail if the economic environment is so unstable that future income is not assured, and thus, risk adjusted return on capital is negative. In such a situation, investors would pay someone (the government) to use the wealth instead to retain at least the value of the principle, or to minimize the erosion of the value relative to when holding it oneself.[6]

ii)  Equivalently, if such a situation prevails, then the value of currently held wealth stock is perceived to be larger in the future than now, i.e., negative discount rate. An agent who tries to maximize consumption over life time will save lessif the net present value of his/her wealth (future consumption) rises.

iii)  On a similar vein, if the expected decline in the size of the labor force is such that it will exceed any realistic productivity increase, then as in i) above, negative interest rate could prevail. However, in an open economy, it is likely to lead to a readjustment of capital flows, i.e., capital inflows into a region with expected increase in labor (or productivity).

iv)   From the market’s point of view, if the current debt stock of the economy is larger than its net present value of total future income, the economy will try to deflate itself out of the debt—whichis an alternative way of taxing households.[7]The reduction of the value of debt can be more instantaneous with the increase in inflation, but in the absence of inflation, negative real interest rate is an alternative way to chipping away the stock of debt, (unless the value of asset is reduced through a crash).

“

Current vs. neutral real rate of interest

Currently, real long term rates in advanced economies are volatile, but along a declining trend(Figure 1). This current episode also differs from the past in that low or negative real interest rates are attributed to the fall in nominal rates and not a sudden rise in high inflation. While it is debatable whether real interest rate should be zero or even negative, based on historic norm, it is on the low side. In fact, real long-term interest rate (defined as the unweighted average of G10 group’s 10-year government debt deflated by CPI)[8] averaged 1½-2%during 1880s and 1990s,or 2½-3% excluding the two war periods (Figure 2). Why the long-term real interest rate was negative during the wars is an issue beyond this paper, but they likely include negative return on capital and were willing to pay a premium for holding government paper (one of the safest way to retain the value of one’s wealth).

Economic history in advanced economies, as captured well in the G10 study, indicate that small positive real interest rate is the “norm” or the market determined neutral rate. More specifically, in the case of the US the average post-war real interest rate (10-year TBs) is around 2½, which is broadly similar to the average of G10 countries.10-year government bond yields fell briefly below zero also during 1974-75, 1979-80, and 2011-12, but largely due to spurts of inflation.

“

 Why neutral real rate of interest may be low

a. Lost opportunity of savingsin the past

Savings in advanced economies peaked around 24% of GDP in the late 1980s, and continued tofall, leveling off at around 20% of GDP after a sharp dip following the global financial crisis (Figure 3). This is broadly comparable to world average savings to PPP weighted GDP over the last 50 years, which was about 22%. The gradual decline since the 2000s is also consistent with theexpected savings behavior and global demography. Figure 4 shows how savings rate would have diverged from the average in Europe and the US if households had smoothed their consumption over the long term(based on the UN projection of global demography).While the accuracy of the estimated magnitude is questionable, at least it provides a rough sense of how savings could have evolvedover time if households were rational, i.e., Europe should have saved most during the 1980s, and start dissaving at around 2020. The US case is broadly similar except that its savings should have peaked during the 2000s.

Real government bond yields in China-Japan-Korea (CJK)[9] have been positive and trending down for the last several decades (Figure 5). Negative real interest rates were present during the first half of the 1970s in Japan and Korea due to high inflation related in part to oil price as well as global monetary conditions. As for China, high inflation that pushed real interest rate down into negative numbers during the late 1980s and mid 1990s were by-products of rapid transition efforts by the government. In more recent years, Japan’s real interest rate fell below zero in line with several other advanced economies. Long term average (1980-16) of about 2½ % is broadly similar to that of the US. Korea’s s real interest was higher at 5%, reflecting a much faster economic growth during this period. In the case of China, it is difficult to make a simple comparison of longer term development due to the transition period as noted earlier. But still, average real interest rate since 2000 is about 3½ % (compared with 2% for Korea) reflecting higher return on capital.

CJK’s actual savings rates (Figure 6) broadly match the demography-based estimation of savings (obtained in the same way as those elaborated for advanced economies earlier; CJK estimates are shown in Figure 7). According to the latter, China’s savings should have peaked during the late 2000s, Japan on the other hand, should have peaked during the early 1990s and Korea during 1990-2000. Unlike in the US and Europe, actual savings broadly mirror these estimations.

Two factors, however, could push China and Korea(and to a lesser extent Japan) to save more even though there was no lost opportunity to save. First is the recent realization of growing longevity well beyond what was expected one or two decades ago. Second is the gloomier global economic outlook without a clear time horizon for a recovery. Since CK economies are largely export dependent, these changing global prospects also mean slower domestic economic growth. Both factors have strong incentive to push up savings, subjecting them to similar situation as the US and the EU. These points are illustrated in the chart belowfor China (Figure 8).

b. Debt and liquidity overhang

“Other liabilities” in Figure 8 include financial liabilities or public debt (or both although they are not additive by nature). Financial liabilities are intra-generational liabilities and under a well- functioning system would not be an issue as one part of the population owes to another part of the same population. It becomes a concern only when the indebted group is highly income constrainedwith debt service payments posing a challenge.Public debt has a similar impact on consumption except that its debt service burden is more widely spread through higher future tax burden. Both financial assets and government debt have risen sharply as share of GDP in the last two decades (Figure 9 and 10 respectively).

Financial assets and liabilities are a way ofholding savings of one group into the future, to be repaid at a later date. If the size of the financial assets becomes too large relative to the production capacity (i.e., expected income stream), thenthere is large probability that the value of savings will erode. This can take place through bursting of an asset bubble, gradually through negative real interest rate, sustained high inflation, orthrough the political process. If the former becomes a systemic risk, it gets converted into the latter as was the case in the US and the EU following global financial crisis.

“

Conclusion: A question of low consumption now or later

If the estimateddemography-based consumption smoothing path during the 1980s-2010s is of any guide, then the rational response of households faced with aging is to catch up, i.e., to save more now relative to the long run average. However, this would mean compression of consumption below the long-term average, discouraging investment. This in turn would suppress growth setting off a vicious cycle. There are two ways to overcome the current low growth trap.

One way to address such a vicious cycle is to push down the real interest rate below the already low neutral rate of interest, possibly into negative rates. This will induce agents to save less. However, if the underlying problem is structural in nature, such a push will only have a temporary effect, and the economy will be in a worse situation in a couple of years, requiring further reduction in the real interest rate. Another way would be to leave the real interest rate as is and lethouseholdsto make up for their lost opportunity to prepare for aging. This would drag down growth in the short run, but once targeted wealth stock is reached, consumption would recover.

It is not clear which option will deliver better results, or minimize the losses in terms of lifetime consumption. However, if the current problem is structural in nature, the former will only postpone the needed reform and nurture a bigger problem. If left alone, the market will try to correct itself and be accompanied by a reduction in public debt and financial assetswith corrective structural reforms setting the stage of a stronger recovery path. In both cases, real interest rate will likely remain below historic average. Yet, pushing it further down into large negative territory through policy rates may risk pushing the real neutral rate down further.

[1] “Saving, Investment and Real Interest Rates” A Study for the Ministers and Governors by the Group of Deputies, IstitutoPoligrafico E ZeccaDelloStato, October 1995

[2]Real interest rates in this note refer to long term government bond yields (10 year maturity except when no data are available) as they are more relevant to real economic activity and entail sovereign risks only.

[3]For more details, please see “Optimal Liquidity and Economic Stability”L. Han and I. Lee, IMF Working Paper, No. 12/135, May 01, 2012

[4] The Political Economy of Silvio Gesell: A Century of Activism; Werner Onken. The American Journal of Economics and Sociology Vol. 59, No. 4 (Oct., 2000), pp. 609-622

[5]The difference of Gesell’s argument from this note is that he limits the use of money for transactions purpose.

[6]This point is similar to Barrow; “Rare Disasters and Asset Markets in the Twentieth Century,” The Quarterly Journal of Economics, August 2006.

[7] Carmen M. Reinhart and M. Belen Sbrancia, “The Liquidation of Government Debt”NBER Working Paper No. 16893March 2011

[8]We use 10 year debt despite the uncertainty added by the term premium because this is the actual measure that affects investment decisions at each point in time.

[9]Deflated by CPI except for China during 1980-85 where annual GDP deflator was used instead. Lending rates are government bonds for Japan and Korea, and lending rates (source: IFS) for China.

观点整理  程曦

图文编辑  程曦

点击查看近期热文

欢迎加入群聊

为了增进与粉丝们的互动，IMI财经观察将建立微信交流群，欢迎大家参与。

入群方法：加群主为微信好友（微信号：imi605）,添加时备注个人姓名（实名认证）、单位、职务等信息，经群主审核后，即可被拉进群。

欢迎读者朋友多多留言与我们交流互动，推荐好文章可联系：邮箱[e-mail]；电话[phone]

关于我们

中国人民大学国际货币研究所（IMI）成立于2009年12月20日，是专注于货币金融理论、政策与战略研究的非营利性学术研究机构和新型专业智库。

中国人民银行副行长潘功胜与诺贝尔经济学奖得主蒙代尔出任IMI顾问委员会主任，委员包括埃德蒙、陈云贤、汉克、李若谷、李扬、马德伦、任志刚、苏宁、王兆星、吴清、夏斌、亚辛·安瓦尔等12位国内外著名经济学家或政策领导人；中国人民银行副行长陈雨露出任学术委员会主任，委员会由47位来自国内外科研院所、政府部门或金融机构的著名专家学者组成。IMI所长、联席所长、执行所长分别为张杰、曹彤、贲圣林，向松祚、涂永红、宋科任副所长。

自2012年起，研究所开始每年定期发布《人民币国际化报告》，重点探讨人民币国际化进程中面临的重大理论与政策问题。报告还被译成英文、日文、韩文、俄文、阿拉伯文等版本并在北京、香港、纽约、法兰克福、伦敦、新加坡和阿拉木图等地发布，引起国内外理论与实务界的广泛关注。

迄今为止，研究所已形成“国际金融”、“宏观经济理论与政策”、“金融科技”、“财富管理”、“金融监管”等五个研究方向，并定期举办货币金融圆桌会议、大金融思想沙龙、燕山论坛、麦金农大讲坛、陶湘国际金融讲堂等具有重要学术影响力的高层次系列论坛或讲座。主要学术产品包括IMI大金融书系、《国际货币评论》（中文月刊&英文季刊）、《IMI研究动态》（周刊）、《国际货币金融每日综述》（日刊）等。

国际货币网：www.imi.org.cn

微信号：IMI财经观察

（点击识别下方二维码关注我们）

只分享最有价值的财经视点

We only share the most valuable financial insights.