A Definitive Guide to China's Economic Cycle
摘要 - 中国短周期的回暖已逐步见顶，而长期的经济处于下行趋势 – 这是我们发现的中国三年库存投资周期模型的结论。数据和图表都清晰地印证了我们三年的投资库存周期理论，并与许多经济文献中论述的其他国家的Kitchin基钦库存周期基本一致。最近，较高频率的数据，如汽车销售等令人失望，零售增长和消费者价格指数疲弱。这些数据侧面证实了我们的模型结果。
我们的模型结论和去年五月份“权威人士”的观点一致，并与市场共识认为中国经济周期“已经见底”，“新的经济周期正在开始”的观点大相径庭。在高库存，高杠杆，和几乎所有资产高估值的大环境里，政策运行的空间日益缩小。随着贫富分化变得越来越根深蒂固，资产泡沫反而已经成为底层群众改变人生的最后一根救命稻草。随着利率被压抑，对更美好生活的渴望正在推动劳动人民群众在一个又一个泡沫里历险折腾。现在，房地产泡沫的风险不容小觑。
此外，上游商品再库存周期和中国周期板块相对表现的势能已经在去年五月份见顶。这些周期变量长期以来持续领先中国股票，中美长债收益率，商品和房地产价格，以及美国周期板块的相对表现。
中国的三年库存投资周期，上游商品再库存周期和中国周期板块的相对表现的势能见顶的情况表明：1）再通胀交易正在逐渐落幕，全球周期性板块将开始跑输大市; 2）大宗商品、债券收益率的上行势能将衰减; 3）中国的房价周期已经见顶，虽然房地产泡沫似乎暂时坚不可摧。我们3月7日发表了题为《再通胀交易已经落幕；准备开启防御性轮动》的报告。最近几个星期以来，上述的这些趋势已经初露端倪。随着防御性轮动在未来几周内逐渐展开，久违了的市场波动性将回归。
这是今天的报告《中国经济周期权威指南》的英文版。中文翻译版随后发。
-----------------------------------
Summary: China’s cyclical upswing is peaking, and its longer-term trend is declining, as our model of 3-year inventory investment cycle suggests. Our 3-year cycle is evident by data and on charts, and is consistent with the Kitchin Inventory Cycle that has been identified in other countries in numerous economic literatures. Recently, higher frequency data, such as disappointing car sales, retail sales growth and a weak CPI, have confirmed our model results anecdotally.
Our findings confirm the writings of the “Authoritative Figure” last May, and contradict the popular belief that China’s economic cycle has “bottomed”, and “a new economic cycle is beginning”. With high inventory, high leverage and high valuation in almost all asset classes, the room for policy maneuvers is limited. As the divide between the rich and poor becomes ever more entrenched, asset bubbles have become one last way to change one’s life. With suppressed interest rates, the desire for a better life is driving the hoi polloi to z-turn from one bubble to another. Now, the property bubble is the elephant in the room.
Further, the strength in upstream commodity restocking cycle and Chinese cyclical’s relative performance have peaked last May. These cycles have demonstrated a consistent long lead over Chinese stocks, long bond yields in China and the US, commodity and property prices, as well as the US cyclical’s relative performance.
The peaking of the 3-year inventory cycle, upstream restocking cycle and Chinese cyclical’s relative performance suggest: 1) the reflation trade is fading soon, and global cyclical will start to underperform; 2) bond yield’s surge will lose momentum; 3) China’s property price cycle has peaked, although the bubble has been proved to be resilient. Indeed, some of these trends have emerged in recent weeks since our last note “The Reflation Trade is Over; Get Set for Defensive Rotation” on March 7. As the defensive rotation unfolds in the coming weeks, volatility will return.
------------------------------
China’s 3-Year Inventory Investment Cycle
A consistent 3-year inventory investment cycle and a long-term down trend. The current cycle has peaked. Consensus believes that China’s economic cycle has bottomed, yet with scant data support. We believe that consensus has mistaken a small cyclical upswing within China’s 3-year inventory investment cycle for the start of a new cycle. And we demonstrate this cycle comprehensively in Focus Chart 1. We can see that China’s current cyclical upswing indeed started from early 2016, coinciding with the inception of the dramatic surge in commodity prices. The start of the upswing predates Trump's winning of the US election. By now, the strength of the upswing appears to have peaked, and will soon decline, if the cycle we have discovered holds. And the strength in commodity price and bond yields should wane.
Further, we note that in the past two decades of which period we have data, the trend of investment growth is down, with lower highs and lower lows in each of the cycle sequentially. This is not difficult to fathom: China’s massive investment scale and increasing leverage are suppressing marginal investment return, and thus limiting the room for further productive investment (Focus Chart 1).
We derive our 3-year investment cycle by measuring the deviation of the actual property investment growth data from its long-term trend. There are four and a half, very clearly defined 3-year cycles: 2003-2006, 2006-2009, 2009-2012, 2012-2015, and last quarter of 2015 till now. We then compare the investment cycle with the rebar price cycle and China’s interest rate cycle. We can demonstrate that these various cycles are closely correlated (Focus Chart 1). That is, the 3-year investment cycle can explain the movements in Chinese asset prices.
Focus Chart 1: The 3-year Inventory investment cycle versus rebar price and long-bond yield.
China’s 3-year cycle is indeed a property inventory investment cycle. Even though China’s cycles are clearly demonstrated in charts, an obvious question begging to be asked is that why the cycle lasts three years? We believe the duration of the cycle is related to the building construction cycle. To construct a 30-storey residential building, the building completion time is around 9-12 months, water and electrical installation around 3 months and plus some more time for safety inspection and miscellaneous approvals. The total time to completion is around 1.5-2 years. Then the building inventory will take up to 1 year to clear, making the building inventory investment cycle to be around 3 years.
As the current investment cycle peaks, the momentum in economic recovery will soon wane, together with the demand for commodity. Bond yield’s strength will weaken, too, as inflation pressure slowly dissipate, and the demand for safe yield in a risk-averse environment grows.
China’s stock market cycle and M1 lead the investment cycle by ~3 months. That said, the relationship between the investment cycle and Chinese stock market is not as tight. Indeed, Chinese stocks tend to lead the investment cycle by around 3 months, although stocks show an inverse relationship with the cycle during the 2015 market bubble (Focus Chart 2). Meanwhile, we can show a more consistent relationship between China M1 narrow money supply growth and the 3-year investment cycle, with a 3-month lead (Focus Chart 2). For equity investors, our task is not only to decipher the cycle within the Chinese economy and its relationship with asset prices, we must also postulate what leads the investment cycle and asset price cycle to earn an edge to trade stocks.
Focus Chart 2: China’s stock market cycle and M1 lead the investment cycle by ~3 months.
The Four Economic Cycles
Cycles can be used to detect inflection points. The term “cycle” suggests regularity, implying that economic variables fluctuate around a long-term trend with well-defined length and amplitude. In reality, of course, economic variables can be fickle and trends can be transient. The eventual realization of a visible trend, after much volatility, can easily outlast the transient memories and the career of many market researchers. And even if trends eventuate, their length and strength will vary considerably.
But these challenges have never stopped economists from their endeavor to decipher economic cycles and trends from data. In fact, the significance of any cyclical models is to identify turning points, especially for market participants. The timing of the cycle’s peak and trough will mean profit or loss for traders. But the reason less so. Usually, doubts arise as there appears to be little theoretical explanations for the alleged cycles, despite clear data evidence.
The four economic cycles of various lengths. There are four cycles. In the order of length, these cycles are: 1) the Kitchin Inventory Cycle (3-5 years); 2) the Juglar Investment Cycle (7-11 years); 3) the Kuznets Building Cycle (15-25 years); and 4) the Kondratieff Basic Capital Goods Cycle. Roughly, 1 Kondratieff = 3 Kuznets = 6 Juglar = 12 Kitchins, as Schumpeter suggested. The cycles are intertwined and interweaved. “Every cycle of higher order may be considered as the trend of the cycle of the next lower order” (Schumpeter, 1939). Upswing and downswing are defined relative to trend. Data can show an uptick for a particular month or quarter within a down trend. The 3-year property inventory investment cycle in China we identify previously is akin to the Kitchin Cycle, and it is peaking.
Note that the accuracy of cycle prediction is inversely correlated with the length of the cycle. The longer the cycle, the more vague and inscrutable the lurid details in the economic data will become. As such, the Kondratieff Cycle is easily the most disputed. Samuelson dismissed the Kondratieff Cycle with a footnote: “whether these long waves are simply historical accidents due to chance gold discoveries, inventions and political wars, it is still too soon to say”.
Focus Chart 3: The long-term cycle in the US economy.
Even for Kondratieff who theorized the cycle, he sounded hesitant as he postulated about the cycle in his first paper: “in the preceding sketch we had no intention of laying the foundations for an appropriate theory of long waves”. At that time, Kondratieff believed that the cycle was endogenous to a capitalist system, and that an upswing was emerging. Such believes contradicted the “communist ideal”, and was deemed “reactionary”. Kondratieff was soon deported to Siberia without trial, and sentenced to solitary confinement.
In Focus Chart 3, our study of the US data in the past 100 years confirms the Kondratieff Wave. For the lack of long time series, we are not able to demonstrate the long waves in the Chinese economy in this report. We are confident that it exists, as it has been demonstrated in many other economies. But such long waves are of little relevance to shorter-term trading calls. Indeed, even the Juglar Cycle and the Kutnets Cycle that tend to last well over a decade, but significantly shorterthan the Kondratieff Cycle, can be imprecise when applied to short-term market fluctuation, and easily outlast many investment analysts’ career. We are a little concerned when we see pundits using the Juglar Cycle to explain strong sales of heavy trucks and excavators in recent quarters. It is like shooting mosquitoes with a canon.
China’s upstream inventory cycle is a long leading indicator.
China’s upstream inventory cycle is a long leading indicator for asset prices. And it has also peaked. Our research shows that China’s upstream inventory cycle is a long leading indicator for asset prices. It leads Chinese stocks by around 9-12 months, and the Chinese property price cycle by even longer (Focus Chart 4). This upstream cycle has been in place consistently for well over a decade, and demonstrated a consistent property of leading the other asset prices (Please refer to our recent report “The Relation Trade is Over; Get Set for Defensive Rotation” on 20170307). Its lead over the other asset classes is even longer than China’s real interest rate cycle, a leading indicator we have discussed for a few times in our previous report. (Please refer to “Sweet and Sour Hog Cycle” on 20160418. This report marked a significant local peak in China’s iron ore and rebar prices).
Focus Chart 4: Cycles in various asset classes in China intertwining across time, with upstream inventory cycle leading
Intuitively, upstream inventory is the most sensitive to the changes in the economy. Such sensitivity gives the upstream inventory cycle its property as a long leading economic indicator. As producers and merchants expect an economic upturn, they express their expectation by replenishing inventory. Their restocking activities can then set the tone for others, and initiate a virtuous cycle as the upstream restocking effects flow through unto mid and down stream industries, and vice versa. If the upstream iron ore cycle maintains its lead by around 9-12 months, and the cycle has peaked around last May, then now stocks strength will start to wane, too.
Focus Chart 5: The strong seasonality in upstream commodity cycle - it has peaked.
Upstream inventory has clearly peaked: We have also found strong seasonality in rebar inventory cycle – it tends to build up during the first quarter of every year, and then gradually depletes for the rest of the year (Focus Chart 5). Importantly, of the ten inventory cycles in the past decade, only 2007, 2009, 2011 and 2016 did not completely deplete the inventory accumulated during the year by year end. Before 2016, these inventory cycles tended to be two years apart, denoting the length of the cycle completion. Each of these heightened inventory cycles in Focus Chart 5 corresponds to the peak years seen in Focus Chart 4. The upstream inventory cycle in 2009 is most unusual, with significant excess inventory left at year end – most likely the aftermath of the 4-trillion stimulus pact.
2016 again experienced a restocking frenzy, after five years of dormancy. Some investors asked whether the 5-year hibernation would mean a stronger and longer restocking cycle. But we note that iron ore inventory at ports has reached new historic high. Even if inventory level can continue to pile up, its speed of rising will decelerate, and the strong momentum in commodities will likely wane consequently (Focus Chart 5).
China’s cyclical relative strength as a long leading indicator.
China’s cyclical relative performance peaked last summer; it suggests defensive rotation now. While the debate about the reflation trade intensifies, China’s cyclical sectors’ relative performance had already peaked last summer – at roughly the same time when the upstream iron ore cycle peaked around May 2016. This observation contradicts the general sense that cyclical sectors have outperformed. Given the close correlation of cyclical sectors’ relative strength with the economic cycle, this observation also suggests that the so-called “cyclical recovery” is probably a chimera, and the momentum in the current recovery is about to roll over.
Focus Chart 6: Relative performance of Chinese cyclical leads China/US 10y and US cyclical relative performance by ~9m.
Surprisingly, our research reveals that the relative performance of China’s cyclical tends to lead US cyclical sectors’ relative strength, and both US and China’s 10-year yield by around nine months (Focus Chart 6). The relative strength in Chinese cyclical over defensive measures the embedded investor expectation regarding growth outlook. It leads ec 66 23964 66 15792 0 0 3943 0 0:00:06 0:00:04 0:00:02 3944onomic growth by around nine months as well. As China’s growth outlook fades, bond yields and foreign cyclical should fall. We believe that the momentum in the current cyclical upswing is waning. And so will the upward thrust in US and China’s 10-year yield.
Hao Hong, CFA
2017-03-23