2018 Hong Kong Stock Investment Strategy Report: Slowly Bull or Back Down?

格隆汇 · Dec 12, 2017 14:19

作者：国信证券王学恒

全球处在一个超级货币宽松的周期（80年代-今）已逾30年。从2000年以后，货币政策不再是主动未雨绸缪而是被动的应对危机。08年金融危机之后全球利率水平已经降无可降（英格兰银行称5000年最低），因此，自2015年11月，美联储开启了历史上时间跨度最长的、艰难的、必要的升级周期。

判断大盘的方向，有两个独立的出发点，一是时间，二是价格。从80年代至今的5轮加息周期来看，在升息周期的中期，往往是股市见到年级别的高点的时间，我们测算平均升息150BP大盘见顶。依此，我们须留意2018年上半年，本轮升息将会走到累计150BP的“敏感区域”。另一方面，我们对A股、港股、美股的盈利增速测算，美股的EPS增速高点是今年2季度，A股是今年3季度，港股是今年4季度（因为港股没有3季报），由于基数效应，三大市场明年EPS的增长率会缓过今年。也就是说，从流动性水平（加息）以及盈利（上市企业EPS增速）两个角度看，三大市场2018年都难利于估值的进一步提升；从估值的角度，目前港股的估值为历史平均水平，也低于美股与A股。地产、汽车、科技（主要是腾讯以及手机产业链）、保险经历了两年的估值修复，龙头公司估值已经显著高于历史平均水平，而银行、原油、电信运营商则低于历史平均水平，后者依旧存在估值修复的空间。

这样，导致了以下的判断：

从2-3年周期来看，港股依旧处在一个估值上修，企业盈利改善的牛市中，它的投资价值显著：1）期间有大量的A/H折价标的；2）像腾讯控股、港交所这类的A/H中独一无二的标的；3）新股IPO，尤其是科技创新类的标的如众安在线、阅文集团、易鑫集团不断的丰实了港股面向新经济的特征；4）以及未来的“创新板”等政策红利；5）国内南下资金占比的不断抬升。因此，港股的未来定位为大陆的“国际版”是充分而恰当的，也是大陆经济全球化与国际化，在金融领域连接世界的旗帜。

而站在1年的角度，我们还需关注估值修复的速度。倘若市场经历短暂的调整之后，2018年上半年以较快的速度完成了对绝大多数板块的估值修复，这就不得不警惕下半年至2019年上半年的回吐压力——它可能来自1）房地产新开工低于预期而带动的上游诸多行业的回落，进而对金融板块估值压制；2）美国加息带动的英国、欧洲央行跟随而带来的资金流出效应；3）部分龙头公司估值较高，EPS增速低于预期；4）国内明年国债收益率振荡下行而阻碍了对银行的息差抬升预期；5）原油价格上涨可能导致的通胀压力。

结合以上两方面意见，我们对港股明年的总体判断是：健康的走法是慢牛徐行，不健康的走法是前扬后抑。然而，在牛市特征中的美股和港股，让我们相对倾向于后一种走势。如果这样，明年上半年，将会是奠定全年收益率水平的关键时刻。

主线上，如果今年的主线是经济复苏，聚焦龙头；明年可能要关注的是均值回归——高估值龙头无法再高，低估值的标的则有回补空间，此外，不应忽视原油价格的波动对诸多行业以及美元汇率的影响。

科技创新方面，我们认为云计算、3D结构光、手机并行运算（NPU）将是2018年的三大方向。云计算大幅提升算力和企业的效率，3D光将连接虚拟与现实，NPU是实现终端AI的基础保证。

1 美联储升息中的股票市场表现

1 港股：A股的基本面，全球的资金面

研究港股市场的难度是，这个市场中存在三类不同的标的：国内企业，香港企业，海外企业；存在三类不同的投资人：海外投资人，香港投资人，大陆投资人。因此，港股在长期的走势中，始终保持了和美股、A股较强的关联性。

从1990年有上证指数开始计算，港股市场与美股、A股的相关性基本相当。在2001-2004，由于恒生指数的内银股较少，港股走的是牛市行情，而A股走的是熊市行情，2005年内银股上市之后，港股与A股的相关性逐渐走强，与美股的相关性逐渐走弱，但港股目前还是与美股的相关性更强：这是因为，港股的投资人海外占比更高，在他们的投资体系中，港股是新兴市场的一员。而在国内资金的眼中，港股是估值更低的A股。因此，海外资金会按照新兴市场的判断来配置港股，而国内资金则是按A/H的联动性来配置港股。

我们以沪港通标的为例。2004年之前，沪港通标的中约有131家非大陆企业，62家大陆企业，大陆企业占比为32%，目前，442家沪港通标的中，大陆企业162家，非大陆企业280家，大陆企业占比提升至36%。

从全市场来看，H股+红筹也在提高。2005年，H股+红筹仅有194只，而今天，这个数量达到了408只，十二年的时间提升了一倍有余。从市值来看，中资股的合计占比为37%（腾讯既不属于H股，也不属于红筹），若加上腾讯，占比约为50%。可以预见的是，随着越来越多的大陆企业赴港IPO，港股未来将会演变为中国证券市场的“国际板”，大陆企业数量、市值在港股的占比将会越来越高。

另一方面，从市场参与者的属地划分来看，2016年中国内地占比21.5%，而排名靠前的地区是英国、美国、欧洲、中国内地、新加坡。中国内地资金占比从2006-2009年前后的8-10%左右，放大到目前的22%左右。

这主要得益于沪港通、深港通的开通，港股通成交占比已经提升至2017年11月的14%左右的水平，一方面，从零到14%，这已经是快速的、巨大的飞跃；另一方面，大陆资金在港股依然处在“影响者”而非“领导者”的地位。或许我们在个别时间、个别板块上对香港市场的走势可以做到了然于胸，但是我们对香港市场的理解可能还远未透彻，因为我们无法尽然把握市场各种资金的动向，它们更多的来自英国、欧洲、美国、新加坡、台湾…我们把港股，近似的比喻成“A股的基本面+海外的资金面” 。

2 抬不起的利率：我们是否还能回到从前？

美国16%的国债收益率水平（1981）已经成为尘封的历史，利率的持续走低，成就了股市的几十年的长牛。美元凭借充当世界货币，通过1）与黄金脱钩；2）多次的量化宽松，实现了对世界财富的聚敛。

从1985年美元指数（兑欧元、日元、英镑、加拿大元、瑞典克朗、瑞士法郎的强弱水平）见164点之后，在长达三十二年的时间里，一浪弱过一浪，这也是美元贬值的直接例证。

当我们把最熟悉的，标普500指数用另一种货币单位——盎司黄金来计量，从1968年美元和黄金脱钩开始，一个或许打破我们认知常识的事实是：标普500并非涨了30多倍，而是下跌了26%！那么这些年，标普500家公司共同成长的过程里，到底给市场赚到“钱”了么？

英格兰银行在 2016 年发布了一个名为“5000 年来最低利率”的观点，阐述了从 3000 年前的美索不达米亚文明的 20%的利率水平到 2016 年美元的不到 1%的利率水平，结论是当下全球利率水平处在 5000 年来最低。

从80年代之后，货币政策总是为经济保驾护航，每一次经济危机，股市大跌，似乎政府，以及政府影响下的美联储，就成了量化宽松的代言人，如果说87年的黑色星期一发生后，美联储还在犹豫中继续保持加息（而在89年才开始缓慢降息），那么2000年的科网泡沫，就成了经济与股市绑架美联储的开始。2000年之后，每一次危机后的半年到一年，都是美联储降息通道的开始，速度一次快过一次，幅度一次高过一次。尽管，在经济恢复时美联储也会小心翼翼的加息，但利息水平似乎更像一个背着沉重包袱的老者，步履蹒跚，一浪低过一浪。

于是，金融市场似乎自反馈式的产生了一个新的“依赖式”的逻辑：经济不好，美联储不会痛快的加息，于是股票上涨；经济好了，上市企业ROE改善，股票继续上涨！似乎，这个新的逻辑中认为，在万般不确定中，最值得信赖的，也是最确定的因素就是美联储——就像容忍孩子犯错的父母一样，她总是呵护市场，不会丢下孩子不管的！

3 美联储升息周期中的美股市场表现

我们认为，目前对香港以及全球市场影响比较大的资金面的潜在利空是全球处在一个升息周期中。因此有必要简要的回顾历史上美股的升息周期带给市场的一些影响。从 80 年代至今，美联储共有 6 轮加息。分别起始于是 1983 年、 1987 年、 1994 年、1998 年、 2003 年、 2015 年。我们将每轮时间跨度、起点终点利率、加息次数列表如下。从幅度上看，每轮加息的利率平均上升水平在 3-4%（除了 2000 年本来升息周期遭遇了科网泡沫而戛然而止）；

从时间跨度上看，本轮将会是时间间隔最长的一次加息周期，我们暂以 2019 年 12 月作为加息终点，本轮加息周期时间跨度将持续 49 个月。正因为如此，本轮股票市场的牛市持续时间已经 8 年之久。

83年首次加息到大盘（标普500）见顶间隔为7个月，期间加息7次，87.5BP，1Y国债高点（收益率低点，下同）时间提前股市9个月，10Y国债高点时间提前股市7个月。加息9次，累计幅度200BP大盘见低点（而不是在加息周期结束之后）。

87年首次加息到大盘（标普500）见顶间隔为3个月，期间加息3次，87.5BP，这里爆发了著名的“黑色星期一”，标普10月19日当天跌幅超过了20%！1Y国债高点时间提前股市12个月，10Y国债高点时间提前股市12个月。由于跌幅过于快速而有幅度，因此加息3次，累计幅度85BP大盘见低点。1Y和10Y国债低点都滞后于股票低点17-18月。

94 年首次加息到大盘（标普 500）见顶几乎同步。 1Y 国债高点时间提前股市16 个月，10Y 国债高点时间提前股市 4 个月。加息 3 次，累计幅度 75BP 后大盘见低点。 1Y 和 10Y 国债低点都滞后于股票低点 8-9 月。

99 年首次加息到大盘（标普 500）见顶间隔为 9 个月，期间加息 5 次，100BP。1Y 国债高点时间提前股市 17 个月，10Y 国债高点时间提前股市 18 个月。加息 6 次，累计幅度 125BP 之后，遭遇了著名的科网泡沫大崩盘，打乱了美联储的加息节奏，也是数次加息周期幅度最小的一次。美联储在 2001 年转向大规模防水，依然未能阻止跌跌不休的股市。 1Y 和 10Y 国债低点提前于股票低点29-33 月。

美国从互联网泡沫的泥沼中走出后，04 年首次加息到大盘（标普 500）见顶间隔为 52 个月，期间加息 17 次，400BP。 1Y 国债高点时间提前股市 52 个月，10Y国债高点时间提前股市53个月。这轮经济复苏的很好，即便加息也未把股市压制，直到加息结束，股市才开始了大溃退——2008年金融危机。

本轮加息起始于2015年12月，分别在2015-12-17日加息25BP、2016-12-15日加息25BP、2017-03-16加息，25BP、2017-06-15日加息25BP。截至2017年11月份，加息4次，合计100BP。

以上，我们总结一下加息对美国股票市场影响：

1、大盘见顶：从几次经验总结，平均加息 6.4 次（平均加息 150BP）之后，大盘见顶。只有一次，在经济增长非常强劲（2004-2006 年）的时候，可以在加息周期中大盘一直涨，当然后来遭遇的回撤是巨大的；而大多数时期，大盘的顶点都是在加息周期的初、中期出现的；

2、国债与大盘：每次都是国债提前大盘见顶（收益率见底），但是提前的时间较难估量；

3、大盘见底：大盘不必要在加息周期结束后见底，它往往在加息周期中后期形成底部，如果大盘的短期跌幅较大，像 87 年的黑色星期一，也可能在加息初期就形成底部。唯有一次，科网泡沫，大盘在加息周期结束后数月才见底。

2 国内：预计2018年GDP增速略降，国债收益率将呈现逐步下降态势

1地产投资：大概率见顶回落

房地产投资的取向是当前市场分歧最大的项目，部分投资者认为房地产投资增速还将继续上行的主要原因在于：库存低以及土地出让情况较好。从土地成交价值累计同比增速来看，其整体仍然处于相对高位。

而从销售增速领先于地产投资的角度出发，基于商品房销售已于2016年3月触顶回落，我们认为房地产投资大概率已经见顶回落。

2 汽车行业：2018年销量增速约为0~3%，低于2017年（4%）

回顾：17 年 1-10 月汽车消费低迷。 2017 年 1-10 月，汽车产销累计分别完成2295.68 万辆和 2292.71 万辆，同比分别增长 4.27%和 4.13%。相比去年而言，2017 年汽车工业增速显著回落，主要受购置税减半政策退坡影响，2017 年消费热情不高（尤其是 1.6L 及以下车型销量增速下滑明显），2017 年全年增速预期在 4%左右，销量或达 2900 辆。

2018 年，在消费提前释放和 2017 年的高基数背景下，销量增速或将持续放缓，但整体仍然维持正增长预期。我们预计 2018 年销量增长区间在 0%-3%，具体增速预期取决于购置税政策的走向，若明年仍然维持 7.5%的购置税，2018 年预期销量增速接近 3%；若明年购置税回升至 10%，2018 年预期销量增速接近 0%。

3 对2018宏观经济判断：稳中有落

对于 2018 年三类重要的经济指标，我们的判断如下：1、 GDP 实际同比增速：2018 年 GDP 同比增速稳中有落，将会在 6.5-6.7%之间；

2、 CPI：居民消费价格指数（CPI）重心走高，2018 年 CPI 中枢平均在 2-2.5%之间，但是环比增速应该弱于季节性；

3、 PPI：工业品价格指数（PPI）具有最大的不确定性，通过情景分析，我们预计 2018 年 PPI 平均水平将在 3-3.5%之间。

纵观 2017 年全年，CPI 的平均增速可能在 1.5%左右。在中性假设条件下，根据我们的预测，2018 年 CPI 的平均水平约为 2.3%，但目前正处于以猪价为代表的食品下行周期中，实际水平大概率弱于中性假设预期，而按照 2 年的下降周期来看，这一回落至少会持续到 2018 年下半年。因此，在考虑猪价等因素后，我们认为 2018年 CPI 中枢平均在 2-2.2%之间，低于中性假设预期。

对于PPI在2018年究竟会处于什么水平，我们可以借鉴两个历史极致乐观水平：

1、如果当期阶段的“供给收缩”效应强于2008前的“需求扩张”效应：这一种情况下，2018年定基指数会到134，那么2018年PPI的平均增速可能会在3.2%左右的水平。

2、如果当期阶段的“供给收缩”效应强于2011年前的“四万亿”效应：2018年定基指数在这一情况会到138，那么2018年的PPI平均增速约为4.5%。

而对于上述两种非常乐观的情景假设来说，真实的水平应该在其基础上下调，因此，基于这一判断，我们预计2018年PPI平均水平在3-3.5%之间。

从历史的统计关系来看，GDP 平减指数与 CPI 和 PPI 的线性关系大致如下：GDP 平减指数=0.8*CPI+0.2*PPI

而根据上述分析，年度来看，若 CPI 从 1.5%的水平回升到 2.2%左右的水平，PPI 从 6.4%的水平下降至 3.5%左右的水平；那么 CPI 的回升将会拉升 GDP平减指数约 0.5 个百分点，而 PPI 的下降将会拉低 GDP 平减指数约 0.6 个百分点。综合来看，根据 CPI 和 PPI 变化合成的 GDP 平减指数增速很可能将于 2018年出现下降，而如果 GDP 实际增速稳中有降，则整体 GDP 名义增速将比 2017年有所回落。

由于，由于 2017 年中国经济复苏强于美国，加之两次加息及金融加强监管等诸多原因，导致了中美息差持续扩大，但我们认为这一趋势将会在 2018 年初终结。原因是：1）国内经济在 2018 年将会增速放缓；2）强金融监管导致了今年的阶段性的利率水平上移；

3 海外：基数效应的压力将于2018年逐步显现

1 高基数压力将于明年显现

由于港股大多数企业不公布三季报，为了审视中国经济自下而上在季度上的盈利同比情况，我们选择了全 A 股 2755 家上市企业利润增速作为辅助观察。由于 2016年，实体经济的基数较低，从 2017 年，以地产、汽车、周期品为代表的行业形成正向拉动，2017 年 Q1 上市公司利润增速达到了 22%，扣除两桶油的影响，同比利润增速为 17%，Q1-Q2 为 17%，Q1-Q3 为 19%，市场预期全年将落在 15%（对已有分析师覆盖的 1825 家企业的预测净利润加总，主要是剔除一些小市值公司与ST），但我们认为该数字可能略微保守，最后的数字大约落在 17%左右。因此，我们倾向性得到如下结论：A 股上市企业利润增速于 Q3 见顶，Q4 的同比增速低于 Q3。如果审视 2016 年的 Q2、 Q3、 Q4 的数据，2017 年 Q1-Q3的增速较快主要是去年的低基数造成的，从 2016 年 Q4 开始，基数抬高，因此2017 年的 Q4 至 2018 年的 Q3 连续四个季度同比压力将逐步显现。

如果分行业来看，材料、能源、工业、电信服务四个行业 2017 年 Q1 净利润增速大幅提高，按照历史平均水平测算，该三个行业将在明年 Q1 开始面临利润增长压力；公用事业、金融今年增速低于历史平均水平，2018 年增长压力不大。信息技术、医疗保健在 2017 年增速逐季度抬高，因此在明年 Q1-Q2 期望能够保持相对的高增长；房地产、可选消费、日常消费为历史平均水平，明年有望保持。简言之：上游行业、中游行业明年一季度增长压力开始显现，下游行业的增长压力在明年下半年。金融与公用事业净利润增长压力最小。

2 美股：比中国的经济恢复更弱一些

美国市场也是如此。在 2017Q1，标普 500 的利润增速为 15%， 2017 年 Q2 为 12%，2017 年 Q3 为 4%。如果说中国企业的利润增速是 Q1-Q3 逐步抬高，美国企业的利润增速则是 Q1-Q3 逐步下降。因此，美国的经济复苏比中国更弱一些。

从各个行业来看，能源和材料改善的较为明显，而影响标普利润较大的几个板块金融的Q3增速放缓明显，其他几个行业—可选消费下滑（迪士尼、福特汽车等）、医疗保健下滑（强生、吉列德科学等）、工业零增长（通用电气、达美航空等），日常消费较弱（沃尔玛、可口可乐等），信息技术表现尚好，在信息技术中，表现最好的二级行业是半导体行业，今年Q2-Q3的增速为100%左右（intel、德州仪器、美光科技、英伟达等）。

3 港股：2017年报将是净利润增速高点

港股的利润增速在 2017H1 实现了 17%的同比增长，参照 Q3 国内企业的利润增速计算，2017H2 还应更胜一筹。

从港股 wind 一级行业净利润增速观察，能源、材料在大宗商品价格强劲反弹的背景下增速较快，工业、房地产因为低基数因素增长强劲，信息技术在腾讯、半导体的带动下也有不俗的表现，医疗保健、金融复苏明显但未超过 2014 年的水平。由于煤炭价格高企，公用事业表现最差。电信运营商与日常消费、可选消费表现一般。按照分析师对覆盖的 474 家公司的前瞻预期，2017 年全年利润增速 24%，而 2018年增速为 13%。也就是说，港股在 2017 年 H2 将达到净利润增速高点，而 2018年 H1、 2018 年 H2 净利润增速将逐季下滑。

4 2018港股大盘：指数先高后低

2006 年-2007 年，恒生综合指数的 PE 水平在 15-20 倍，PB 水平在 2-3 倍，尽管港股自 2016 年涨到目前，PE 水平仅为 12.7，PB 水平仅为 1.3。因此，与历史比较的恒生估值水平仅是平均水平。从长期讲，港股处在一个更大的牛市里：1、中国经济处在复苏中，周期企业的盈利有利于上游企业降杠杆，进而修复从上游到银行的资产负债表，使得金融企业的估值长期抬高，有助于提升大盘估值水平；2、龙头公司的集中，如保险、地产、汽车等行业的集中，有利于抬升指数；3、港股较 A 股的估值折价，随着国内资金的不断南下，以及未来资金面的进一步放开（如早年曾提到的 QDII），也有利于抬升市场估值。

但是，我们不认为这个过程将发生在明年。主要几点考量是：

1、全球的升息周期结束大概率落在 2019 年底-2020 年中。由于美 10 年期国债收益率并未形成向上突破，但随着加息进入中后期，向上突破（如 2.8-3%）又是早晚的事情，而美股8年、港股 2 年的牛市将会 2018-2019 年的某个时间终结而转向一个数月级别的调整；

2、美股、 A 股、港股明年的 EPS 增速都会有所下降，2018 年国内由房地产、汽车拉动力量大幅减弱，进而对上游周期品行业的需求下降，使得经济在同比上表现不如今年；

3、目前的估值水平，尤其是龙头公司的估值水平已经在市场预期的 2018-2019年水平，体现、部分体现了 2018-2019 年的增长。

如果我们拉出腾讯、碧桂园、吉利汽车、舜宇光学科技、瑞声科技、中国平安、建设银行、汇丰控股、中国石油、中国移动的估值水平，我们会发现：对于今年财报靓丽的公司：如腾讯、碧桂园、吉利汽车、舜宇光学科技、瑞声科技，它们的市值高点的估值水平都已经逼近或者超过了历史高位。尽管最近价格有所回调，但如果这些标的在 2018 年上半年反弹较快，则随着 2018 年 H1 的整体财报的增速压力，它们在全球升息周期中继续抬高估值难度较大。

另一方面，银行、原油、电信运营商、公用事业板块处在历史的中低水平，对港股大盘又起到稳定剂的作用。考虑到 2017 年 H2 将会是港股企业 EPS 增速的高点，且目前在升息周期初期，这是通常股市估值快速抬升的时候：企业盈利改善，资金面配合；而年报季过后，我们认为伴随着地产行业的新开工同比持平或下滑，汽车销量可能略低于今年的可能，带动上游的盈利增速放缓，加之升息周期拉动长期国债收益率上行，将大概率终结估值的上修，进而股市将进入估值平稳或下降的调整期。而另一方面，由于龙头公司在 2017 年的估值溢价充分，我们认为这种趋势在 2018年全年将不可持续。相反，一些增速高，估值低的中小市值标的，超跌之后，因其较好的性价比，将会受到南下资金和投资人的青睐。但值得注意的是，这些中小市值标的对大盘指数的影响是较小的。综上，我们认为明年港股大盘有两种走法：慢牛徐行 VS 前扬后抑。健康的走法是慢牛徐行，不健康的走法是前扬后抑。然而，在牛市特征中的美股和港股，让我们相对倾向于后一种走势。如果这样，明年上半年，将会是奠定全年收益率水平的关键时刻。

4 2018年投资建议：看好金融、原油、电信、医药，创新关注云计算、 3D 结构光

对比恒生一级行业的 10 年估值水平，目前低于历史平均估值水平的板块分别是能源、电讯、公用事业、金融、综合。

1 原油：价格稳定向上，抬升板块估值

历史经验数据显示，从长期趋势来看，原油价格的趋势性变化，在两次石油危机以后都是由需求而非供给驱动的。从 1986 年至 2002 年，原油的价格在近 20 年的时间里维持了区间震荡的走势，而这一时期原油的价格走势基本上完全由西方发达国家（G7 国家）的工业增速决定。 2003 年以后，中国经济快速崛起，打破了之前原油的区间走势。 2003 年至今的原油走势基本上是由中国的工业和投资增速决定的。目前全球经济处于复苏上行通道，原油需求有望继续上升，原油价格在未来一段时间将持续保持震荡上行趋势。

作为上游资源品，原油用途极为广泛，原油价格的变动通过产业链条可以将其影响传导至多个行业，对各行业的影响也不尽相同。

石油开采业是油价上涨的最直接受益者；而煤炭作为石油替代品可间接得益于国际油价走高对能源价格的推动作用；石油加工、炼焦及核燃料加工业主要产品包括汽油、柴油、煤油等成品油及化工原料等石油制品，其作为相对上游行业也将受益于原油价格上涨。原油价格的上涨可能压缩橡胶及塑料制品行业、纺织服装行业、航空运输业等下游行业的利润空间。原油价格的上涨直接推高了橡胶及塑料制品行业及纺织服装行业的成本。航空运输业因航油占民航业成本四分之一以上，油价上升直接导致营运成本及费用大幅提升。

随着原油价格的复苏，将有助于上市公司的估值修复。相对而言，采掘占比更高的公司对原油价格的波动更加敏感，潜在的弹性更强。

2金融：行业的估值水平将进一步修复，推荐银行、保险板块

几年前，对银行业拖累较重是上游企业，主要是煤炭、化工、钢铁等行业，伴随这些行业的供给侧改革以及大宗商品的复苏盈利，我们认为银行业的坏账率将会被有效控制在合理范围并进一步下降。此外，由于制造业和零售业，主要是一些中小企业的不良贷款比例较高，因此，只要中国经济不遭遇严重通缩，我们认为银行业的不良率会逐渐降低。国有商业银行、股份制商业银行、城市商业银行、农村商业银行中，国有商业银行的不良率已经连续六个季度下降，这时我们认为国有商业银行的 PB 可以进一步修复的基础。

3 电信：运营商盈利能力大幅转好，中国移动的稳定性更佳

进入 2017 年，电信运营商的收入增速明显较 2015 年、 2016 年有恢复，且新增移动电话用户数也明显高过 15 年、 16 年。

由于担心移动互联网对电信运营商的增值业务冲击，5G 的投入增加电信运营商的资本开支，2015-2016 年电信运营商的估值处在历史底部。中国联通混改后，2017年相对表现亮眼的标的只有中国联通（0762.HK），而中国电信（0728.HK）、中国移动（0941.HK）表现平平。我们认为，2018 年电信运营商将开始 5G 试点，但不会大规模的投放 5G 设备，主要原因是：1）国家尚未有颁发 5G 牌照的预期；2）全球 5G 技术还在不断优化中；3）尚未有成熟的、规模的支持 5G 标准的手机芯片。因此，我们认为三大运营商 2018 年的资本开支将保持在 3000 亿元左右。综上，资本开支下降，收入增长，估值水平处在历史底部，将有利于该板块的表现。

中国的电信运营商运营更稳健，如中国移动的有息负债率为 0，电信 14%，联通30%，而美国的 AT&T、 VERIZON、 T-MOBILE 分别达到了 34%、 46%与 42%，因此，在升息周期的背景下，美国电信运营商的表现将弱于中国的电信运营商。

从估值来看，EV/EBITDA 比率国内运营商具有明显优势（中国移动 5.4，中国联通 4.9，中国电信 3.6）

4 医药：关注最具创新优势和长线价值的医药港股

政策颠覆——重塑医药产业竞争格局。最近两年多，医药行业各类政策频出，政策引导对行业未来发展产生深远的影响。整体来看，产业将实现优胜劣汰的升级，集中度上升。

医药经济——健康回升。 据工信部统计数据，医药工业主营收入增速由 2015年的个位数回升至 12.1%，营业利润增速由 2015 年的 12.2%回升至 18.4%，营业利润率也较最低点不断回升。 SW 医药生物板块上市药企主营收入增速由2014、 2015 年的低点回升至 20%左右。 SW 医药生物板块上市药企归母净利润 2017 年前三季度增速达到 22.6%，为近年来归母净利润增速高点。

差距即机遇，我国生物制药产业将迎来黄金发展期。 2015 年，全球生物制药规模已突破 2500 亿美元。 2016 年，我国医药工业主营业务收入中生物药品制造达到3350 亿元（2010-16 年生物药品 CAGR 24%vs 医药工业整体 CAGR16%）。但对比国际市场，我国生物药规模目前占比仍然较低（我国 8%vs 全球 25%）。随着 2015年-2020 年国际多款重磅生物药面临专利悬崖，我国生物制药将迎来黄金发展期。他山之石——创新造就巨头。美国安进公司通过自身研发和并购不断推出重磅生物制药产品，自 1984 年上市以来股价已累计上涨 1100 倍，市值屡创新高，超过千亿美元。目前我国生物创新药企业大展宏图的环境正逐渐完善，有望向安进等国际巨头看齐。

2017 年，A 股和 H 股市场医药板块市值 TOP10 公司大多取得远超行业平均的收益。我们认为从长期来看，随着政策推动下的产业升级，优势龙头企业有望长期受益产业集中度的提升甚至有望引领产业发展，但 2018 年若海外利率水平上升较快的话，短期可能估值继续抬升的情况较难。

关注创新药龙头公司：我国药品创新土壤逐渐肥沃，创新药将迎来黄金发展期，重点推荐业绩增长明确、研发投入大、在研药物储备丰富且 2018 年估值低于 30 倍的大市值龙头公司：石药集团（1093.HK）、复星医药（2196.HK）、三生制药（1530.HK），这类企业主要靠稳健的成长来维持估值或者对冲有可能的估值阶段性的、小幅度的回撤。

高成长性的 H 股稀缺标的：高成长性、未来潜力巨大的 H 股稀缺投资标的：CRO龙头药明生物（2269.HK）、 CAR-T 技术领先的金斯瑞生物科技（1548.HK）。这类公司是我们中长期跟踪的标的，因为它们市场空间巨大且在新产品盈利释放时爆发力最强；重点推荐医药商业龙头公司：医药商业长期来看将在政策助力下不断提升集中度，龙头企业长期受益于行业整合。短期来看，包括上药在内龙头企业受调拨业务影响医药商业部分业绩增速放缓。预计明年中期之后影响有望减弱，增长有望提速，重点推荐估值较低且长期受益于行业整合的全国性商业龙头公司：国药控股（1099.HK）、华润医药（3320.HK）和上海医药（2607.HK）。

重点推荐估值偏低、成长性确定的细分领域龙头公司：显著低估的中药配方颗粒龙头：中国中药（0570.HK）、受益行业深度整合，业绩快速弹升的大输液龙头：石四药集团（2005.HK）、核心产品高增长，在研产品潜力大的抗流感化药龙头东阳光药（1558.HK）

5 电子：3D 光学与 NPU 是 2018 年的创新元年

3D 成像是手机端又一次革命性升级。 3D 成像关键点在于获取深度信息，实现拍照技术上从二维到三维的跨越，是手机端 2018-2019 的创新的焦点。目前实现的手段有三种：（1）结构光，代表公司及产品有苹果 iPhone X（收购该领域最大公司 Prime Sense）、微软 Kinect-1 等；（2）双目可见光，代表公司产品Leap Motion；（3）飞行时间法（TOF），代表公司产品微软 Kinect-2。

三种方案中，双目方案最为复杂，算法要求很高，而 TOF 和结构光两种方案优势在于算法简单、响应速度快，识别距离大等优势。 TOF 与结构光相比，结构光在集成度、成本、成像速度方面占据优势，结构光更适合用于前置摄像头，TOF 方案抗干扰性更好，探测距离远，适合用于后置镜头。

以苹果新一代内嵌应用动话表情为代表，通过获取人脸的三维数据，可以得到人面部表情的喜怒哀乐，从而可以使 3D 表情做出与人相类似的表情，与 5G 高速率结合，提升前置摄像头可玩性。人脸识别作为继指纹识别后的又一大生物识别技术，因其独特的方便、经济及准确性而越来越受到世人的瞩目。然而传统人脸识别往往简单通过前置摄像头获取的二维图像进行识别，由于数据量较少，其安全性并不高，例如采用上述技术的三星 Galaxy Note8，使用照片就能骗过人脸识别功能。借助 3D sensing 系统，iphone X 可以绘制出独一无二的3D 面谱。相比于二维图像，iphone X 采集到的人体面部 3D 数据能够分辨出的几何精度会明显提高，可以大大降低了错误率。不仅如此，该功能的速度会比Touch ID 指纹识别的速度更快。

3D成像助力手机端AR/VR腾飞。信息的获取是AR/VR的核心也是必经之路，3D成像有助于AR/VR迅速打开手机端应用市场。与2D成像相比，3D成像在硬件上新增了红外光源与红外摄像头，是手机产业链上全新技术，将使提前布局厂商深度受益。

3D成像使人脸识别替代指纹识别。指纹识别需要人主动输入信息，且正面需要较大留白，与全面屏趋势不符。与指纹识别方案相比，3D成像方案则更加自然安全，人脸面向手机时即可实现解锁，且不占据手机正面过多位置，在全面屏时代具有更多优势。

3D成像打开手机产业链零部件新增长空间。根据智研咨询预测，到2022年3D成像技术将率先在消费电子应用上大规模落地，随后逐步辐射到其他行业，到2022年，3D成像在消费电子领域市场规模将达到60亿美元以上，2016-2022年复合增长率高达160%。对于产业链而言，3D成像带来诸多新增零组件，在发射端，新增了红外激光发射器和辅助元件；在接收端，除了可见光镜头外，还新增红外接收部分，包括镜头、红外传感器和窄带红外滤光片，而图像处理器IC在算法上也会带来相应提升，也将比单纯的2D成像IC成本更高。3D成像能大幅增加产业链价值，给红外核心厂商带来高业绩弹性。

瑞声科技（2018.HK）推出了晶圆级玻璃镜头(MLL)。它的优势是：1）MLL 方案具有更小的尺寸，有助于 3D 遥感相机模组装入全屏智能手机的窄边；2）更优异的光学特性。瑞声科技混合镜头（玻璃+塑料）已进入客户检验程序。 3）拥有更低的成本。两英寸的晶片可被切割成 100 片以上的镜头，与传统玻璃镜头生产流程相比更助于降低成本。 4）避免潜在的专利纠纷。由于瑞声科技布局光学较早，在 2009 年先后投资了 Heptagon、欧洲 Kaleido、 I.Square 等企业，成为其光学技术的主要来源。其中，公司收购 Kaleido 获得了晶圆级玻璃镜头制造能力。此外，其在国内外建设了 6 个光学设计中心，专注发展创新光学设计。

超高清迎来最好的时代。4G时代一般下载速度都超过15mbps。智能手机用户可以凭此在线听歌看视频。按照5G的规划，运营商很有可能把手机互联网网速提高到10gbps或更高水平，大概是现有网速的100倍，比现有固网还会快。高清电影数秒就能下载完成，即便是边走边下也没问题。这必将带来移动视频业务的进一步增长。4G时代手机网络视频用户规模及占比均飞速增长，截止2016年底，手机网络视频用户规模接近5亿人，移动网络视频用户占手机网民比例达91.8%。5G到来提高网速的同时，硬件的支撑也在进步，与3D成像VR/AR相配合，未来网络视频有望实现更多玩法。

从市场需求看，目前人工智能芯片的需求可归纳为三个类别：一是面向于各大人工智能企业及实验室研发阶段的训练环节市场；二是数据中心推断，亚马逊Alexa等主流人工智能应用，均需要通过云端提供服务，即推断环节放在云端而非用户设备上；三是面向智能手机、智能安防摄像头、机器人/无人机、自动驾驶、VR等设备的设备端推断市场，设备端推断市场需要高度定制化、低功耗的人工智能芯片产品。如华为在Mate 10的麒麟970中搭载寒武纪IP，旨在为手机端实现较强的深度学习本地端计算能力，从而支撑以往需要云端计算的人工智能应用

从芯片的分类看，人工智能芯片经过 CPU、 GPU、 FPGA 的尝试，寒武纪人工智能 ASIC 芯片已经问世，这是人工智能发展的重大转折点，是从硬件底层的芯片为人工智能发展提供了动力，其效果远超过先进算法。但是，人工智能类 ASIC 的发展仍处于早期。其根本原因是，ASIC 一旦设计制造完成后电路就固定了，只能微调，无法大改。而硬件的研发设计与生产成本很高，如果应用场景是否为真市场尚不清晰，企业很难贸然尝试。此外，能设计出适用于人工智能芯片的公司必然是要既具备人工智能算法又擅长芯片研发的公司，进入门槛较高。

理想中的人工智能芯片需要具备：

1.可编程性：适应算法的演进和应用的多样性；2.架构的动态可变性：适应不同的算法，实现高效计算；3.高效的架构变换能力：< 10 Clock cycle，降低开销;4.高计算效率：避免使用指令这类低效率的架构。5.高能量效率：~5TOps/W，功耗 < 1mW ，识别速度 > 25f/s。6.低成本：能够进入家电和消费类电子；7.体积小：能够装载在移动设备上；8应用开发简便：不需要芯片识别方面的知识。

按照以上条件，目前的CPU、CPU+GPU、CPU+FPGA、CPU+ASIC的作法均不是理想的架构。

依据清华大学微电子研究所的研究成果，可重构计算芯片最适合人工智能。可重构芯片支持软件硬件编程，硬件随软件变化而变化，也就是软件定义芯片。软件的规模可以任意大，但硬件的规模总是有限的，所以（1）需要将软件划分为与硬件规模相适应的模块，逐块运行（2）硬件结构必须动态可变以适配每个软件模块的功能。

所以如果把可重构计算与经典计算进行比较，经典结构是刚性的，而可重构计算是函数化的柔性的。在性能上偏重于 ASIC，灵活性上偏重于处理器。两者兼而有之。

虽然可重构芯片处于概念提出阶段，离真正落地还有一段距离，但是这为我们研究人工智能芯片提出一种视角，有可能成为人工智能的转折点。

6 计算机：云计算、云服务是 2018 年的主旋律

从成本优势看，目前中国人均 GDP 已经是美国人均 GDP 的 1/6，而十年前，该数字为 1/20，如果人均 GDP 可以视为人均收入的替代，则中国的软件行业劳动力成本的大幅抬高，让行业不得不思考利用云计算、 SAAS 等模式大幅度提高劳动生产率，降低人头式的项目制开支。

此外，美国正是人均 GDP 达到万美元级别，第三产业占比达到 55-60%（60 年代末，大约相当于目前中国的水平），信息化的革命浪潮（计算机、互联网、云计算）不断的打破传统，提高生产率，造就了微软、甲骨文、亚马逊、 Salesforce 等一批软件企业的几十年的长牛。麦肯锡认为，决定 2025 年经济的 12 大颠覆技术除了移动互联网，还有知识工作的自动化（这是最贴切的对 SAAS 的描述），物联网，云计算，先进机器人，自动汽车，基因组，储能技术，3D 打印，先进油气开采，先进材料和可再生能源。我们看到，SAAS 与云计算几乎占据了榜单中非常靠前的位置。

2017 年，国内 SAAS 市场的规模为 6 亿美金，预计未来以年化 25%的速度增长（全球 CAGR 为 17%）。在公有云市场方面，中国在 2017 年市场空间将达到 30 亿美元，其中主要部分为 IAAS（以阿里为代表），其次 SAAS 与 PAAS 各 6 亿美元。

云计算已成为数据中心行业之主要增长动力。云计算使用托管于互联网之远程伺服器网络（而非本地服务器或个人计算机）以存储、管理及处理数据之常规做法。现今几乎所有消费者数据应用均作为云服务交付，及企业应用数据亦正快速转移至云计算。于 2019 年之前，超过 80%工作量将由云数据中心处理；而余下部分则将由传统数据中心处理。云计算需要大量物理传感器接插件、路由器、电力、信号及网络，从而需要大量的光学传感器与连接器。因此，我们认为，除了 SAAS、软件企业，云计算的上游，光器件、光芯片、连接器等会直接受益于数据中心的投资。

国内的中际旭创、光讯科技，港股的鸿腾精密（6088.HK，收购 avago 的光模组部分），昂纳科技集团（0877.HK）是这个大市场的主要玩家。伴随国产 100G芯片的推出，中国的光芯片与光器件，将在全球提升竞争力，并有望将能力延展到工业机器视觉、汽车 ADAS 等领域。

7 互联网：吃鸡游戏有望成为 2018 年的“LOL”级产品

12 月 1 日，腾讯正式宣布将代理今年全球火爆的《绝地求生》，并公布了《绝地求生》端游国服端专属运营计划和生态共建战略等规划。《绝地求生》是蓝洞（Bluehole）和 H1Z1 等游戏的制作人 Playerun known 联合开发的大逃杀类游戏，每局游戏有最多 100 名玩家参与，可以选择单人、双人、四人组队三种模式。游戏开始时所有玩家都会被空投到一个绝地岛上，岛上会有每隔一段时间逐渐缩小的安全区域，区域外会有毒气造成毒气圈内玩家体力值下降直到“阵亡”。

因此所有玩家都需要尽可能在被毒气侵袭前进入安全区域，并在沿途和偶遇的其他玩家决一胜负。游戏最终将只有一个人/团队可以获得胜利，胜利者屏幕上会出现“大吉大利，晚上吃鸡”（Winner winner, chicken dinner）的字样。

自 2017 年 3 月上市至今，《绝地求生》全球累计售出超过 2280 万份，连续 38周以上蝉联 steam 销量冠军。目前《绝地求生》同时在线玩家人数超过 200 万人，远高于其后的 dota2 和 csgo。目前占游戏玩家总数超三分之一的国内玩家愿意忍受高延迟和不稳定的服务器，支付每月 30-60 元的加速器费用来玩这款售价 98 元的游戏，足见《绝地求生》的受欢迎程度。我们认为随着国服的开启，国内现有 600 万以上的玩家绝大部分将会回归，更多由于网络、设备、服务器等问题还在观望的玩家也会加入，《绝地求生》的市场空间巨大。

前所未有的游戏体验有望支撑moba后的下一个爆款游戏类型。“大逃杀”类游戏是一个开放世界体系，其模式可以带给玩家极为丰富的游戏体验，我们简单总结如下：

未知激发强代入感和好奇心。《绝地求生》的基本规则简单清晰——一定会缩小的安全区，尽力存活更久；但设定（随机的航路、降落地点、捡到物品、安全区的位置）的随机性和参与玩家行为的随机性给游戏带来了巨大的偶然性和不可预知性，将“与人斗和与天斗”结合到一起，极大地提升了玩家的代入感和好奇心。

玩家能收获超高的成就感和兴奋度。每局游戏参与玩家多达100人，若能成功“吃鸡”，意味着战胜了99/98/96个对手，艰巨的挑战能够激发玩家的专注度和求生欲。游戏的过程要经历一无所有——捡装备——逃离毒圈——玩家对抗等多个环节，期间的隐忍、击杀、逃毒，都会给予玩家丰富的游戏体验，提高对付出努力后兑现回报的期待。随着游戏进程的深入，玩家的兴奋度会不断提高，到最后决胜阶段甚至会出现肾上腺素飙升，取胜后的兴奋感和成就感均高于绝大多数场景下的moba和FPS类游戏。

入门门槛低，更易吸纳新玩家。传统的竞技类游戏，如英雄联盟、dota2、csgo等，均是已知队友对手数目，在熟悉的地图和线路上进行对抗，长期的积累演化出较为固定的策略和较高的入门门槛。绝地求生通过多元化的策略选择和很大的随机性，弱化了射击技巧在游戏中的重要性，降低了新手的入门门槛，更容易吸引新玩家甚至是女性玩家的加入。

游戏模式丰富，社交属性明显。《绝地求生》可以选择单人模式、双人模式、四人模式，满足不同玩家需求。游戏中一键队内语音有助于好友开黑、陌生人通过游戏结交，“吃鸡”一词的流行亦体现出这款游戏的社交属性非常明显。

打击外挂和构建生态将是腾讯代理的看点。射击类游戏大量数据运算需要在本地计算机进行，因此外挂难以被彻底杜绝。《绝地求生》的火爆也导致国内出现大批外挂制作者和使用者，严重破坏了广大普通玩家的游戏体验，导致steam上10月以来关于游戏的差评数目飙升。腾讯有丰富的射击类游戏运营经验和先进的大数据检索、分析和处理能力，在代理国服之后能否加大打击外挂的力度，为玩家营造良好的游戏环境，并围绕游戏、直播、赛事等板块打造出完整的生态圈，将会在很大程度上决定这款游戏的寿命。

端游盈利参考steam现有模式。我们认为腾讯代理国服后，《绝地求生》在国内仍大概率沿用其在steam上卖游戏+饰品开箱+交易提成的盈利模式。目前游戏的价格为98元/套，钥匙15元/个，steam会对市场上的交易进行15%的抽成。当前《绝地求生》的服饰、箱子在steam市场上交易火爆，未来若腾讯能根据中国玩家爱好推出特色饰品，亦有望延续这一热度。

移动端“吃鸡”大幕拉开，用户增速迅猛。近期，网易、腾讯小米等公司陆续推出了多款移动端的“类吃鸡”游戏，玩家数目迅猛增长，其中最为火爆的网易《荒野行动》注册用户已经突破1亿，按此DAU突破2000万，创下国内手游DAU超千万的最快记录。移动端吃鸡类手游数目和玩家规模的快速增长反映出吃鸡类游戏对注重休闲和碎片化时间的手游玩家亦具有极大吸引力。

目前网易凭借《荒野行动》和《终结者2：审判日》两款游戏在吃鸡类手游的竞赛中取得了领先位置。腾讯即将推出两款同蓝洞合作的正版《绝地求生》手游：《绝地求生：刺激战场》和《绝地求生：全军出击》，加上此前根据CF类手游改编的《荒岛特训》和评价不及预期的《光荣使命》，腾讯手中的吃鸡类手游将达到4款。我们认为“正版”《绝地求生》手游将是腾讯最大的卖点。凭借《绝地求生》积累的人气和腾讯在社交平台上的龙头地位，若其两款游戏能实现高质量、高还原度等目标，将有望实现后发制人，再造《王者荣耀》级别的爆款手游。

8 部分公司一览

Author: Wang Xueheng of Guoxin Securities

The world has been in a super-monetary easing cycle (1980s-present) for more than 30 years. Since 2000, monetary policy is no longer to take the initiative to prepare in advance, but to respond to the crisis passively. Global interest rates have fallen indefinitely since the 2008 financial crisis (the Bank of England says it is the lowest in 5,000 years), so since November 2015, the Fed has launched the longest, difficult and necessary cycle of upgrading in history.

To judge the direction of the market, there are two independent starting points, one is time, the other is price. From the point of view of the five rounds of interest rate hikes since the 1980s, in the middle of the interest rate hike cycle, it is often the time when the stock market sees other high points in the grade, and we estimate that the average interest rate hike 150BP market peaked. In view of this, we should note that in the first half of 2018, this round of interest rate hikes will reach the "sensitive area" of cumulative 150BP. On the other hand, when we estimate the profit growth rate of A shares, Hong Kong stocks and US stocks, the highest EPS growth rate of US stocks is in the second quarter of this year, A shares are in the third quarter of this year, and Hong Kong stocks are in the fourth quarter of this year (because Hong Kong stocks do not have three quarterly reports). Due to the base effect, the growth rate of EPS in the three major markets next year will be slower than this year. In other words, in terms of liquidity level (interest rate hike) and earnings (EPS growth rate of listed companies), it is difficult for the three major markets to further improve their valuations in 2018; from a valuation point of view, Hong Kong stocks are currently valued at historical averages and lower than US stocks and A-shares. After two years of valuation repair in real estate, automobile, technology (mainly Tencent and mobile phone industry chain) and insurance, the valuation of leading companies has been significantly higher than the historical average, while banks, crude oil, and telecom operators are lower than the historical average. the latter still has room for valuation repair.

This leads to the following judgment:

From the point of view of the 2-3 cycle, Hong Kong stocks are still in a bull market with upward valuation and improved corporate profits, and its investment value is significant: 1) there are a large number of discount targets of Amax H during the period; 2) unique targets of AGreg H, such as Tencent and HKEx. 3) IPO of new shares, especially the targets of scientific and technological innovation, such as Zhongan Online, China Literature, Yixin Group, continue to enrich the characteristics of Hong Kong stocks facing the new economy; 4) and the future "innovation board" and other policy dividends; 5) the rising proportion of domestic southward capital. Therefore, it is sufficient and appropriate for Hong Kong stocks to be positioned as the "international version" of the mainland in the future, and it is also the banner of the economic globalization and internationalization of the mainland and connecting the world in the financial field.

From an one-year point of view, we also need to pay attention to the speed of valuation repair. If the market undergoes a brief adjustment and the valuation repair of most sectors is completed quickly in the first half of 2018, we have to be wary of the pullback pressure from the second half of the year to the first half of 2019-it may come from 1) the decline in many upstream industries driven by lower-than-expected new real estate starts, thus suppressing the valuation of the financial sector. 2) the capital outflow effect of the United Kingdom and the European Central Bank driven by interest rate hikes in the United States; 3) the higher valuations of some leading companies and the lower-than-expected EPS growth rate; 4) the downward oscillation of domestic treasury bond yields next year, which hinders the expectation of higher interest spreads on banks; 5) the inflationary pressure that may be caused by the rise in crude oil prices.

Combined with the above two views, our overall judgment on Hong Kong stocks next year is: the healthy way is slow, and the unhealthy way is to promote the future. However, the US and Hong Kong stocks in the bull market characteristics make us relatively inclined to the latter trend. If so, the first half of next year will be a critical moment to set the level of full-year yields.

On the main line, if this year's main line is economic recovery, focus on the leader; next year may be focused on the mean return-the leader of high valuation can no longer be high, the target of low valuation has room for compensation, in addition, the impact of fluctuations in crude oil prices on many industries and the exchange rate of the US dollar should not be ignored.

In terms of scientific and technological innovation, we believe that cloud computing, 3D structured light and mobile parallel computing (NPU) will be the three major directions in 2018. Cloud computing greatly improves computing power and the efficiency of enterprises. 3D light will connect virtual and reality. NPU is the basic guarantee for the realization of terminal AI.

1 the performance of the stock market during the Fed's interest rate hike

1 Hong Kong stocks: the fundamentals of A shares and the global capital side

The difficulty of studying the Hong Kong stock market is that there are three different types of targets in this market: domestic enterprises, Hong Kong enterprises and overseas enterprises, and there are three different types of investors: overseas investors, Hong Kong investors and mainland investors. Therefore, in the long-term trend, Hong Kong stocks have always maintained a strong correlation with US stocks and A shares.

Since the beginning of the Shanghai Composite Index in 1990, the correlation between the Hong Kong stock market and US stocks and A-shares is basically the same. In 2001-2004, due to the small number of domestic bank stocks in the Hang Seng Index, Hong Kong stocks were in a bull market, while A shares were in a bear market. after the listing of silver stocks in 2005, the correlation between Hong Kong stocks and A shares gradually strengthened, and the correlation with US stocks gradually weakened. however, Hong Kong stocks are still more relevant to US stocks: this is because Hong Kong stocks account for a higher proportion of overseas investors, and Hong Kong stocks are members of emerging markets in their investment system. In the eyes of domestic capital, Hong Kong stocks are A-shares with lower valuations. Therefore, overseas funds will allocate Hong Kong stocks in accordance with the judgment of emerging markets, while domestic funds will allocate Hong Kong stocks according to the linkage of Acord H.

Let's take the Shanghai-Hong Kong Stock Exchange as an example. Before 2004, there were about 131 non-mainland enterprises and 62 mainland enterprises in the Shanghai-Hong Kong Stock Connect, accounting for 32% of the total. At present, of the 442 Shanghai-Hong Kong Stock Connect, there are 162 mainland enterprises and 280 non-mainland enterprises, and the proportion of mainland enterprises has increased to 36%.

From the point of view of the whole market, H shares + red chips are also on the rise. In 2005, there were only 194 H-shares plus red chips. Today, the number has reached 408, more than doubling in 12 years. In terms of market capitalization, Chinese shares account for 37 per cent of the total (Tencent is neither H-share nor red-chip), or about 50 per cent with Tencent. It can be predicted that with more and more mainland enterprises coming to Hong Kong for IPO, Hong Kong stocks will evolve into the "international board" of China's securities market in the future, and the proportion of mainland enterprises and market capitalization in Hong Kong stocks will become higher and higher.

On the other hand, in terms of the territorial division of market participants, mainland China accounted for 21.5% in 2016, while the top regions were the United Kingdom, the United States, Europe, mainland China and Singapore. The proportion of mainland Chinese capital has magnified from about 8-10 per cent around 2006-2009 to about 22 per cent now.

This is mainly due to the opening of the Shanghai-Hong Kong Stock Connect and the Shenzhen-Hong Kong Stock Connect. The proportion of Hong Kong stock transactions has risen to about 14% in November 2017. On the one hand, from zero to 14%, this is already a rapid and huge leap; on the other hand, mainland funds are still in the position of "influencers" rather than "leaders" in Hong Kong stocks. Perhaps we can have a clear understanding of the trend of the Hong Kong market at individual times and in individual sectors, but our understanding of the Hong Kong market may be far from thorough, because we cannot fully grasp the movements of all kinds of funds in the market. They are more from the United Kingdom, Europe, the United States, Singapore, Taiwan … We liken Hong Kong stocks to "A-share fundamentals + overseas capital".

2 unbearable interest rates: can we go back to the past?

The US Treasury yield level of 16% (1981) has become a dusty history, and the continued decline in interest rates has made the stock market a bull for decades. By acting as the world currency, the US dollar has achieved the accumulation of world wealth through 1) decoupling from gold and 2) repeated quantitative easing.

Since 1985, the dollar index (the strength level of the euro, Japanese yen, pound sterling, Canadian dollar, Swedish kronor and Swiss franc) has seen 164points, and over the past 32 years, it has weakened one wave after another, which is also a direct example of the depreciation of the US dollar.

When we measure the most familiar S & P 500 index in another currency unit, ounce gold, since the dollar was decoupled from gold in 1968, a fact that may break our common sense is that the S & P 500 is not up more than 30 times. It's down 26%! So have the S & P 500 companies made "money" for the market in the process of growing together over the years?

In 2016, the Bank of England released a view called "the lowest interest rate in 5000 years", which went from 20% in Mesopotamian civilization before 3000 to less than 1% in 2016. The conclusion is that global interest rates are now at their lowest level in 5000 years.

Since the 1980s, monetary policy has always protected the economy. Every time the economic crisis and the stock market plummeted, it seems that the government, and the Federal Reserve under the influence of the government, have become the spokesmen of quantitative easing. If after Black Monday in 1987, the Fed continued to raise interest rates hesitantly (but only began to cut interest rates slowly in 1989), then the science and technology bubble in 2000 It became the beginning of the economy and stock market kidnapping the Federal Reserve. After 2000, six months to a year after each crisis, it was the beginning of the Fed's interest rate cut channel, faster and higher each time. Although the Fed will raise interest rates cautiously when the economy recovers, the level of interest rates seems to be more like an old man with a heavy burden, hobbling and falling below one wave after another.

As a result, the financial market seems to have produced a new "dependent" logic: if the economy is bad, the Fed will not raise interest rates happily, so stocks will rise; when the economy is good, ROE of listed companies will improve, and stocks will continue to rise! It seems that this new logic holds that the most trustworthy and certain factor among all uncertainties is the Fed-like parents who tolerate their children's mistakes, she always takes care of the market and will not leave their children behind!

(3) the performance of the US stock market in the interest rate raising cycle of the Federal Reserve

We believe that the potential negative side of the capital side, which currently has a great impact on the Hong Kong and global markets, is that the world is in a cycle of raising interest rates. Therefore, it is necessary to briefly review some of the historical impact of the interest rate rise cycle of US stocks on the market. Since the 1980s, the Fed has raised interest rates in six rounds. It began in 1983, 1987, 1994, 1998, 2003 and 2015. We list the time span of each round, the interest rate at the beginning and the end point, and the number of interest rate increases as follows. In terms of range, the average rate rise in each round of interest rate hikes is 3-4% (except for the abrupt end of the interest rate hike cycle in 2000 when it was hit by the science and technology bubble).

In terms of time span, this round will be the longest interest rate hike cycle, we temporarily take December 2019 as the end of the interest rate hike cycle, the current interest rate hike cycle will last 49 months. Because of this, the bull market in the stock market has lasted for eight years.

The interval between the first interest rate hike to the peak of the S & P 500 in 1983 was 7 months, during which the interest rate was raised 7 times. The high point of 87.5 BPJ 1Y treasury bonds (yield low, the same below) was 9 months earlier than that of the stock market, and the high time of 10Y treasury bonds was 7 months earlier. Nine interest rate hikes, the cumulative range of 200BP market saw a low (not after the end of the rate hike cycle).

In 1987, the first interest rate hike to the market (S & P 500) peaked in three months, during which three interest rate hikes occurred. 87.5BP, the famous "Black Monday" broke out here, and S & P fell more than 20% on October 19th! The peak time of 1Y treasury bonds is 12 months ahead of the stock market, and that of 10Y treasury bonds is 12 months earlier. As a result of the decline is too fast and a range, so raise interest rates three times, the cumulative range of 85BP market saw a low. The lows of 1Y and 10Y treasury bonds lag behind the stock lows of 17-18 months.

The first rate hike in 1994 to the market (S & P 500) peaked almost at the same time. The peak time of 1Y treasury bonds is 16 months ahead of the stock market, and that of 10Y treasury bonds is 4 months ahead of schedule. After raising interest rates for 3 times, the cumulative range reached a low point after 75BP. The lows of 1Y and 10Y treasury bonds both lag behind the stock lows in August-September.

In 1999, the first interest rate hike to the market (S & P 500) peaked in 9 months, during which it raised interest rates five times, 100BP. The peak time of 1Y treasury bonds is 17 months ahead of the stock market, and that of 10Y treasury bonds is 18 months ahead of schedule. After raising interest rates six times, the cumulative range of 125BP was met by the collapse of the famous science and technology bubble, which disrupted the pace of interest rate hikes of the Federal Reserve and was the smallest in several interest rate hikes. The Fed turned to massive waterproofing in 2001 and still failed to stop the falling stock market. The lows of 1Y and 10Y treasury bonds are 29-33 months earlier than the stock lows.

After emerging from the quagmire of the dotcom bubble, the United States raised interest rates for the first time in 2004 to the peak of the market (S & P 500) in 52 months, during which it raised interest rates 17 times, 400BP. The peak time of 1Y treasury bonds and 10Y treasury bonds are 52 months and 53 months respectively. This round of economic recovery is so good that even the interest rate hike did not suppress the stock market, and it was not until the end of the interest rate hike that the stock market began to collapse-the 2008 financial crisis.

This round of interest rate hikes began in December 2015, raising interest rates on 2015-12-17 25BP, 2016-12-15 25BP, 2017-03-16, 25BP, 2017-06-15 25BP respectively. By November 2017, interest rates had been raised four times, totaling 100BP.

Above, let's summarize the impact of interest rate hikes on the US stock market:

1. The market peaked: summing up from several experiences, the market peaked after raising interest rates by an average of 6.4times (average 150BP). Only once, when economic growth was very strong (2004-2006), the market could keep rising during the interest rate hike cycle, and of course the subsequent pullback was huge; and for most of the period, the peak of the market occurred in the early and middle of the interest rate hike cycle.

2. National debt and market: every time, the national debt reaches its peak in advance (the yield reaches the bottom), but the time ahead is difficult to measure.

3. The market bottomed out: it is not necessary for the market to bottom after the end of the rate-raising cycle. It tends to form a bottom in the middle and later stages of the rate-raising cycle. If the short-term decline of the market is large, such as Black Monday in 1987, it may also form a bottom in the early stage of interest rate hikes. Only once, the market bottomed out a few months after the end of the interest rate hike cycle.

2 domestic: the growth rate of GDP is expected to slow slightly in 2018, and the yield of treasury bonds will show a trend of gradual decline.

(1) Real estate investment: the probability is that it will peak and fall.

The orientation of real estate investment is the most divergent project in the current market, and some investors believe that the main reason why the growth rate of real estate investment will continue to rise is that inventory is low and land sales are good. From the cumulative year-on-year growth rate of land transaction value, its overall is still at a relatively high level.

From the point of view that the growth rate of sales is ahead of real estate investment, based on the fact that commercial housing sales peaked in March 2016, we believe that the probability of real estate investment has peaked.

2 Automotive industry: sales growth in 2018 was about 0.3%, lower than in 2017 (4%)

Review: automobile consumption was depressed from January to October in 17 years. From January to October 2017, automobile production and sales totaled 2295.68 million and 22.9271 million respectively, an increase of 4.27% and 4.13% respectively over the same period last year. Compared with last year, the growth rate of the automobile industry slowed significantly in 2017, mainly affected by the policy of halving purchase tax. Consumer enthusiasm was not high in 2017 (especially the growth rate of models below 1.6L dropped significantly), and the growth rate for the whole year of 2017 is expected to be about 4%. Sales may reach 2900 vehicles.

Sales growth may continue to slow in 2018 against the backdrop of early release of consumption and a high base in 2017, but overall growth is expected to remain positive. We expect sales growth in 2018 to be 3%. The specific growth rate is expected to depend on the trend of the purchase tax policy. If the purchase tax is maintained at 7.5% next year, the expected sales growth in 2018 will be close to 3%. If the purchase tax returns to 10% next year, the expected sales growth in 2018 is close to 0%.

Three macroeconomic judgments on 2018: there are ups and downs in stability.

For the three important economic indicators in 2018, our judgments are as follows: 1. The actual year-on-year growth rate of GDP: the year-on-year growth rate of GDP in 2018 will be between 6.5% and 6.7%.

2. CPI: the center of gravity of the consumer price index (CPI) is higher, with the CPI center averaging 2-2.5% in 2018, but the month-on-month growth rate should be weaker than seasonal.

3. PPI: the industrial product price index (PPI) has the greatest uncertainty. Through scenario analysis, we expect the average level of PPI to be between 3-3.5% in 2018.

Throughout 2017, the average CPI growth rate is likely to be around 1.5%. Under the neutral assumption, according to our forecast, the average level of CPI in 2018 is about 2.3%, but it is currently in a food downward cycle represented by pig prices, and the actual level is likely to be weaker than expected in the neutral hypothesis, and according to the two-year decline cycle, this decline will continue until at least the second half of 2018. Therefore, after taking into account factors such as pig prices, we believe that the average CPI hub in 2018 is between 2-2.2%, which is lower than the neutral hypothesis.

We can learn from two historical levels of extreme optimism about where PPI will be in 2018:

1. If the "supply contraction" effect in the current period is stronger than the "demand expansion" effect before 2008: in this case, the fixed base index will reach 134in 2018, then the average growth rate of PPI in 2018 may be about 3.2%.

2. If the "supply contraction" effect of the current period is stronger than the "4 trillion" effect before 2011: the fixed base index will reach 138 in 2018, then the average PPI growth rate in 2018 will be about 4.5 per cent.

For the above two very optimistic scenario assumptions, the real level should be lowered on their basis, so based on this judgment, we expect the average PPI level to be between 3 and 3.5 per cent in 2018.

From the historical statistical relationship, the linear relationship between GDP deflator and CPI and PPI is roughly as follows: GDP deflator = 0.8*CPI+0.2*PPI

According to the above analysis, if CPI rebounds from 1.5 per cent to 2.2 per cent and PPI falls from 6.4 per cent to 3.5 per cent, then the rebound in CPI will increase the GDP deflator by about 0.5 percentage points, while the decline in PPI will drag down the GDP deflator by about 0.6 percentage points. Taken together, the growth rate of the GDP deflator based on CPI and PPI changes is likely to decline in 2018, while if the real GDP growth rate slows steadily, the overall nominal GDP growth rate will be lower than in 2017.

Due to the fact that China's economic recovery is stronger than that of the United States in 2017, coupled with two interest rate hikes and strengthened financial regulation, the interest rate gap between China and the United States continues to widen, but we believe that this trend will end in early 2018. The reasons are: 1) the domestic economy will slow down in 2018; 2) strong financial regulation has led to a phased rise in interest rates this year.

3 overseas: the pressure of base effect will gradually emerge in 2018.

1 High base pressure will appear next year

Since most Hong Kong companies do not publish three-quarter reports, in order to examine the bottom-up quarterly earnings of the Chinese economy, we selected 2755 A-share listed companies for profit growth as an auxiliary observation. Due to the low base of the real economy in 2016, industries represented by real estate, automobiles and cyclical products have formed a positive pull since 2017. the profit growth rate of Q1 listed companies reached 22% in 2017, excluding the influence of two barrels of oil. year-on-year profit growth is 17% Q1, Q2 is 17%, Q1, Q3 is 19%, and the market is expected to fall to 15% for the whole year (the forecast net profit of 1825 companies already covered by analysts. Mainly excluding some small market capitalization companies and ST), but we think the figure may be a little conservative, the final figure is about 17%. Therefore, we tend to draw the following conclusion: the profit growth rate of A-share listed companies is the highest in Q3, and the year-on-year growth rate of Q4 is lower than that of Q3. If you look at the data of Q2, Q3 and Q4 in 2016, the faster growth of Q1-Q3 in 2017 is mainly due to the low base last year. Starting from Q4 in 2016, the base has risen, so Q3 in 2017 and 2018 will gradually show year-on-year pressure for four consecutive quarters.

If we look at the industry, the net profit growth rate of Q1 in the four industries of materials, energy, industry and telecommunications services increased significantly in 2017. according to the historical average, the three industries will face the pressure of profit growth in Q1 next year; the growth rate of public utilities and finance this year is lower than the historical average, and there is little pressure for growth in 2018. The growth rates of information technology and health care are rising quarter by quarter in 2017, so Q1-Q2 is expected to maintain relatively high growth next year; real estate, optional consumption, and daily consumption are historical averages and are expected to remain so next year. In short: the growth pressure of the upstream industry and the middle industry began to appear in the first quarter of next year, and the growth pressure of the downstream industry in the second half of next year. The pressure of net profit growth of finance and public utilities is the least.

Us stocks: weaker than China's economic recovery

The same is true of the American market. At 2017Q1, profit growth for the S & P 500 was 15 per cent, with Q2 at 12 per cent in 2017 and Q3 at 4 per cent in 2017. If the profit growth rate of Chinese enterprises is a gradual increase in Q1-Q3, the profit growth rate of American enterprises is a gradual decline in Q1-Q3. As a result, the economic recovery in the United States is weaker than in China.

From the perspective of various industries, the improvement in energy and materials is more obvious, while the Q3 growth rate of finance in several sectors that affect S & P profits has slowed down obviously. several other industries-decline in optional consumption (Walt Disney Company, Ford, etc.), health care decline (Johnson & Johnson, Gilead Science, etc.), zero industrial growth (General Electric Co, Delta Airlines, etc.), daily consumption is weak (Walmart Inc, Coca-Cola Company, etc.) The performance of information technology is still good. In information technology, the best performing secondary industry is the semiconductor industry, with Q2-Q3 growth of about 100% this year (intel, Texas Instruments Inc, Micron Technology Inc, NVIDIA Corp, etc.).

3 Hong Kong stocks: 2017 report will be the highest growth rate of net profit

The profit growth of Hong Kong stocks achieved 17% year-on-year growth in 2017H1, according to the profit growth rate of Q3 domestic enterprises, 2017H2 should be better.

From the observation of the net profit growth rate of Hong Kong stock wind level I industry, energy and materials are growing rapidly against the background of a strong rebound in commodity prices, industry and real estate are growing strongly due to low base factors, information technology is also doing well driven by Tencent and semiconductors, and the recovery in health care and finance is obvious but does not exceed the level of 2014. Because of high coal prices, public utilities are the worst performers. The performance of telecom operators is average with daily consumption and optional consumption. According to analysts' forecasts for the 474 companies covered, profits will grow by 24% for the whole of 2017, compared with 13% in 2018. In other words, Hong Kong stocks will reach the highest net profit growth rate in 2017, while the net profit growth rate of H1 in 2018 and H2 in 2018 will decline quarter by quarter.

4 2018 Hong Kong stock market: the index is high first and then low

From 2006 to 2007, the Hang Seng Composite Index had a PE level of 15-20 times and a PB level of 2-3 times, although Hong Kong stocks have risen since 2016, with a PE level of 12.7 and a PB level of 1.3. As a result, the Hang Seng valuation compared with history is only the average. In the long run, Hong Kong stocks are in a bigger bull market: 1. China's economy is recovering, and the profits of cyclical enterprises are conducive to the reduction of leverage by upstream enterprises, thus repairing the balance sheets from upstream to banks, which makes the valuation of financial enterprises higher for a long time, which helps to raise the valuation level of the market. 2. The concentration of leading companies, such as insurance, real estate, automobile and other industries, is conducive to driving up the index. 3. The valuation discount of Hong Kong stocks compared with A shares, with the continuous southward flow of domestic funds and the further liberalization of capital in the future (such as the QDII mentioned earlier), are also conducive to raising market valuations.

However, we don't think this process will happen next year. The main considerations are:

1. The end of the global interest rate hike cycle is likely to end from the end of 2019 to mid-2020. As the US 10-year Treasury yield has not formed an upward breakthrough, but as the interest rate increase enters the middle and later stage, it is only a matter of time before the upward breakthrough (such as 2.8-3%). On the other hand, the bull market of US stocks for eight years and Hong Kong stocks for two years will end sometime in 2018-2019 and turn to a multi-month adjustment.

2. The EPS growth rate of US stocks, A shares and Hong Kong stocks will all decline next year. In 2018, the domestic driving force driven by real estate and cars will be greatly weakened, resulting in a decline in demand for the upstream cyclical products industry, which makes the economy perform worse than this year.

3. The current valuation level, especially that of the leading companies, has been at the market expected level of 2018-2019, reflecting and partly reflecting the growth of 2018-2019.

If we pull out the valuation levels of Tencent, Country Garden Holdings, Geely Automobile, Sunny Optical Technology, AAC Technologies Holdings Inc., Ping An Insurance, China Construction Bank Corporation, HSBC Holdings PLC, Petrochina and China Mobile Limited, we will find that for companies with beautiful results this year, such as Tencent, Country Garden Holdings, Geely Automobile, Sunny Optical Technology, AAC Technologies Holdings Inc., their valuation levels of high market capitalization are close to or exceed historical highs. Despite the recent price correction, if these targets rebounded quickly in the first half of 2018, it will be difficult for them to continue to raise valuations in the global interest rate hike cycle as the overall earnings report of H1 in 2018 comes under pressure.

On the other hand, banks, crude oil, telecom operators and public utilities are at historical low and medium levels, which play a stabilizing role in the Hong Kong stock market. Considering that H2 will be the highest EPS growth rate of Hong Kong stock companies in 2017, and at the beginning of the interest rate hike cycle, this is usually the time when stock market valuations rise rapidly: corporate profits improve and capital is coordinated. After the annual report season, we believe that with the new construction of the real estate industry flat or declining compared with the same period last year, car sales may be slightly lower than this year's possibility, leading to a slowdown in profit growth upstream, coupled with the interest rate rise cycle pulling up long-term treasury bond yields, which will most likely end the upward revision of valuation, and then the stock market will enter an adjustment period of stable or declining valuation. On the other hand, due to the full valuation premium of leading companies in 2017, we believe that this trend will not be sustainable for the whole of 2018. On the contrary, some small and medium-sized market capitalization targets with high growth rate and low valuation will be favored by southward funds and investors because of their better performance-to-price ratio. However, it is worth noting that these small and medium-sized market capitalization targets have little impact on the market index. To sum up, we think that there are two ways to move the Hong Kong stock market next year: slow Niu Xu Xing VS up and down. The healthy way to walk is to walk slowly, while the unhealthy way to walk is to carry forward and suppress later. However, the US and Hong Kong stocks in the bull market characteristics make us relatively inclined to the latter trend. If so, the first half of next year will be a critical moment to set the level of full-year yields.

4 2018 investment advice: optimistic about finance, crude oil, telecommunications, medicine, innovation focus on cloud computing, 3D structured light

Compared with the 10-year valuation of Hang Seng's primary industry, the sectors currently lower than the historical average are energy, telecommunications, public utilities, finance and integration.

1 crude oil: the price is stable and upward, raising the valuation of the plate

Historical empirical data show that from the long-term trend, the trend changes in crude oil prices are driven by demand rather than supply after the two oil crises. From 1986 to 2002, the price of crude oil maintained the trend of interval fluctuation in the past 20 years, and the price trend of crude oil during this period was basically determined by the industrial growth rate of western developed countries (G7 countries). After 2003, the rapid rise of China's economy broke the previous trend of crude oil. The trend of crude oil since 2003 is basically determined by the growth rate of China's industry and investment. At present, the global economy is in an upward channel of recovery, crude oil demand is expected to continue to rise, crude oil prices will continue to maintain a volatile upward trend for some time to come.

As an upstream resource product, crude oil has a wide range of uses, and the change of crude oil price can transmit its impact to many industries through the industrial chain, and the impact on different industries is also different.

The oil mining industry is the most direct beneficiary of rising oil prices, while coal, as a substitute for oil, can indirectly benefit from the promotion of energy prices by higher international oil prices; the main products of the petroleum processing, coking and nuclear fuel processing industries, including refined oil such as gasoline, diesel and kerosene, and petroleum products such as chemical raw materials, will also benefit from the rising price of crude oil as a relatively upstream industry. The rise in crude oil prices may reduce the profit margins of downstream industries such as rubber and plastic products industry, textile and clothing industry, and air transport industry. The rise in crude oil prices has directly pushed up the costs of the rubber and plastic products industry and the textile and clothing industry. In the air transport industry, aviation oil accounts for more than 1/4 of the cost of the civil aviation industry, and the rise in oil prices has directly led to a substantial increase in operating costs and expenses.

With the recovery of crude oil prices, it will help to repair the valuations of listed companies. Relatively speaking, companies with a higher proportion of extraction are more sensitive to fluctuations in crude oil prices and potentially more resilient.

2 Finance: the valuation level of the industry will be further repaired, and banks and insurance sectors are recommended.

A few years ago, the heavy drag on the banking industry was the upstream enterprises, mainly coal, chemical, steel and other industries, along with the supply-side reforms of these industries and the recovery of commodities profitability. we believe that the bad debt rate of the banking industry will be effectively controlled within a reasonable range and further reduced. In addition, due to the high proportion of non-performing loans in the manufacturing and retail industries, mainly some small and medium-sized enterprises, as long as the Chinese economy does not experience severe deflation, we think that the non-performing rate of the banking industry will gradually decrease. In state-owned commercial banks, joint-stock commercial banks, urban commercial banks and rural commercial banks, the defect rate of state-owned commercial banks has declined for six consecutive quarters. At this time, we think that the PB of state-owned commercial banks can be further repaired.

3 Telecom: the profitability of operators has greatly improved, and China Mobile Limited's stability is better.

In 2017, the revenue growth rate of telecom operators is significantly higher than that in 2015 and 2016, and the number of new mobile phone users is also significantly higher than that in 15 and 16 years.

Worried about the impact of mobile Internet on the value-added services of telecom operators, the investment of 5G increases the capital expenditure of telecom operators, and the valuation of telecom operators in 2015-2016 is at an all-time low. After China Unicom's mixed reform, only China Unicom (0762.HK) was the relatively outstanding target in 2017, while China Telecom Corporation (0728.HK) and China Mobile Limited (0941.HK) were mediocre. We believe that telecom operators will start the 5G pilot in 2018, but will not launch 5G devices on a large scale. The main reasons are: 1) the country has not yet expected to issue a 5G license; 2) the global 5G technology is still being optimized; 3) there is not yet a mature and large-scale mobile phone chip that supports the 5G standard. Therefore, we believe that the capital expenditure of the three major operators will remain at around 300 billion yuan in 2018. Taken together, falling capital spending, income growth and historical lows in valuations will help the sector's performance.

China's telecom operators are more robust, such as China Mobile Limited's interest-bearing debt ratio of 0, telecom 14% and Unicom 30%, while AT&T Inc, Verizon Communications Inc and T-MOBILE of the United States have reached 34%, 46% and 42%, respectively. Therefore, in the context of the interest rate hike cycle, the performance of US telecom operators will be weaker than that of Chinese telecom operators.

From the valuation point of view, the EV/EBITDA ratio of domestic operators has obvious advantages (China Mobile Limited 5.4,4.9 China Unicom, China Telecom Corporation 3.6)

4 Pharmaceutical: pay attention to the pharmaceutical Hong Kong stocks with the most innovative advantages and long-term value

Policy subversion-reshaping the competition pattern of the pharmaceutical industry. In the past two years, various policies have appeared frequently in the pharmaceutical industry, and policy guidance has a far-reaching impact on the future development of the industry. On the whole, the industry will achieve the upgrading of survival of the fittest, and the degree of concentration will increase.

Pharmaceutical economy-- health rebounds.According to statistics from the Ministry of Industry and Information Technology, the growth rate of the main income of the pharmaceutical industry has rebounded to 12.1% from the single digits in 2015, the growth rate of operating profit has rebounded to 18.4% from 12.2% in 2015, and the operating profit margin has also rebounded from the lowest point. The growth rate of the main income of listed pharmaceutical companies in the SW pharmaceutical and biological sector rebounded to about 20 per cent from the lows of 2014 and 2015. The net profit of listed pharmaceutical companies in the SW pharmaceutical and biological sector grew by 22.6% in the first three quarters of 2017, the highest growth rate in recent years.

The gap is the opportunity, and China's biopharmaceutical industry will usher in a golden period of development. In 2015, the global scale of biopharmaceuticals has exceeded 250 billion US dollars. In 2010, the main business income of China's pharmaceutical industry reached 335 billion yuan in biopharmaceutical manufacturing (2010-16 years biopharmaceutical CAGR16% as a whole). However, compared with the international market, the proportion of biomedicine scale in China is still low (25% of 8%vs in the world). With a number of international blockbuster biopharmaceuticals facing the patent cliff from 2015 to 2020, China's biopharmaceutical industry will usher in a golden period of development. The stone of other mountains-innovation makes a giant. Amgen Inc of the United States continues to launch blockbuster biopharmaceutical products through its own research and development and mergers and acquisitions. Since its listing in 1984, its share price has risen 1100 times, and its market capitalization has repeatedly reached record highs, exceeding 100 billion US dollars. At present, the environment for the great development of China's biological innovative pharmaceutical enterprises is gradually improving, and it is expected to be on a par with international giants such as Amgen Inc.

In 2017, the market capitalization of the pharmaceutical sector of the A-share and H-share markets was mostly TOP10, which far exceeded the industry average. We believe that in the long run, with the policy-driven industrial upgrading, dominant leading enterprises are expected to benefit from the long-term increase in industrial concentration and even lead industrial development, but if overseas interest rates rise rapidly in 2018, it is more difficult for valuations to continue to rise in the short term.

Focus on innovative drug leading companies:The soil of pharmaceutical innovation in China is gradually fertile, and innovative drugs will usher in a golden period of development. We will focus on recommending leading companies with clear performance growth, large R & D investment, rich drug reserves under research and valuation of less than 30 times in 2018: CSPC Pharmaceutical (1093.HK), Shanghai Fosun Pharmaceutical (2196.HK) and 1530.HK (Sansheng Pharmaceutical). Such enterprises mainly rely on steady growth to maintain valuations or hedge against possible valuation periodic, small pullbacks.

High-growth H-shares are scarce:High-growth, future potential H-share scarce investment targets: CRO leader Wuxi Biologics (2269.HK), CAR-T technology leader Genscript Biotech Corporation (1548.HK). Such companies are the target of our medium-and long-term tracking because they have huge market space and are most explosive when new product profits are released.Focus on recommending leading pharmaceutical companies:In the long run, the pharmaceutical business will continue to increase the degree of concentration with the help of policies, and the leading enterprises will benefit from industry integration for a long time. In the short term, leading enterprises, including Shangyao, are affected by the transfer business and the growth rate of the performance of the pharmaceutical business has slowed down. The impact is expected to weaken and growth is expected to accelerate after the middle of next year, focusing on national business leaders with low valuations that have long benefited from industry consolidation: Sinopharm Group, 3320.HK and 2607.HK.

Focus on the recommendation of low valuation, growth determined segment of the leading companies: significantly undervalued traditional Chinese medicine formula granule leader:China traditional Chinese Medicine (0570.HK), the big infusion leader with the deep integration of the beneficiary industry and the rapid rise in performance: stone four Pharmaceutical Group (2005.HK), high growth of core products and anti-influenza drug leader Dongyang Guangyao (1558.HK) with great product potential under development.

5 Electronics: 3D Optics and NPU is the first year of Innovation in 2018

3D imaging is another revolutionary upgrade for mobile phones. The key point of 3D imaging is to obtain depth information and realize the leap from two-dimensional to three-dimensional photography technology, which is the focus of the innovation of mobile phone 2018-2019. At present, there are three means to achieve: (1) structured light, representing the company and products include Apple Inc iPhone X (acquisition of the largest company in this field, Prime Sense), Microsoft Corp Kinect-1, etc.; (2) binocular visible light, representing the company's product Leap Motion; (3) time-of-flight method (TOF), and representing the company's product Microsoft Corp Kinect-2.

Among the three schemes, binocular scheme is the most complex and demanding algorithm, while TOF and structured light schemes have the advantages of simple algorithm, fast response and large recognition distance. Compared with structured light, TOF has advantages in integration, cost and imaging speed. Structured light is more suitable for front camera. TOF scheme has better anti-interference and long detection distance, so it is suitable for rear lens.

Represented by Apple Inc's new generation of embedded mobile expressions, the joys and sorrows of human facial expressions can be obtained by obtaining 3D facial data, which can make 3D expressions similar to human expressions and combine with 5G high-speed expressions. improve the playability of the front camera. As another major biometric technology after fingerprint recognition, face recognition has attracted more and more attention because of its unique convenience, economy and accuracy. However, traditional face recognition is often carried out simply through the two-dimensional image obtained by the front camera, due to the small amount of data, its security is not high. For example, Samsung Galaxy Note8, which uses the above technology, can deceive the face recognition function by using photos. With the help of the 3D sensing system, iphone X can draw a unique 3D face spectrum. Compared with two-dimensional images, the geometric accuracy of human facial 3D data collected by iphone X can be significantly improved, and the error rate can be greatly reduced. Not only that, this function will be faster than Touch ID fingerprint recognition.

3D imaging helps mobile AR/VR take off. The acquisition of information is the core and the only way of AR/VR. 3D imaging helps AR/VR to quickly open the mobile application market. Compared with 2D imaging, 3D imaging adds infrared light source and infrared camera to the hardware, which is a new technology in the mobile phone industry chain and will benefit manufacturers in advance.

3D imaging enables face recognition instead of fingerprint recognition. Fingerprint identification requires people to actively input information, and the front needs a large amount of white space, which is not in line with the trend of the overall screen. Compared with the fingerprint recognition scheme, the 3D imaging scheme is more natural and secure, can be unlocked when the face is facing the mobile phone, and does not occupy too many positions on the front of the phone, so it has more advantages in the era of full screen.

3D imaging opens a new space for the growth of components in the mobile phone industry chain. According to Zhi Research Consulting, 3D imaging technology will be the first to land on a large scale in consumer electronics applications by 2022, and then gradually radiate to other industries. By 2022, the market size of 3D imaging in the consumer electronics sector will reach more than 6 billion US dollars, with a compound growth rate of 160 per cent from 2016 to 2022. For the industry chain, 3D imaging brings a lot of new components, infrared laser emitters and auxiliary components are added at the transmitter; at the receiver, in addition to the visible lens, infrared receiver is also added, including lenses, infrared sensors and narrowband infrared filters, and the image processor IC will also bring corresponding improvement in algorithm, which will be more expensive than the simple 2D imaging IC. 3D imaging can greatly increase the value of the industry chain and bring high performance flexibility to the infrared core manufacturers.

AAC Technologies Holdings Inc. (2018.HK) launched a wafer-level glass lens (MLL). Its advantages are: 1) the smaller size of the MLL scheme, which helps the 3D remote sensing camera module fit into the narrow edge of the full-screen smartphone; 2) better optical properties. AAC Technologies Holdings Inc. mixed lens (glass + plastic) has entered the customer inspection process. 3) have lower cost. A two-inch wafer can be cut into more than 100 lenses, helping to reduce costs compared to the traditional glass lens production process. 4) avoid potential patent disputes. Due to AAC Technologies Holdings Inc. 's early layout optics, he successively invested in Heptagon, European Kaleido, I.Square and other enterprises in 2009, which has become the main source of its optical technology. Among them, the company acquired the wafer-level glass lens manufacturing capacity by acquiring Kaleido. In addition, it has built six optical design centers at home and abroad, focusing on the development of innovative optical design.

Ultra high definition ushered in the best times. In the 4G era, download speeds are generally faster than those of 15mbps. Smartphone users can listen to music and watch videos online. According to 5G's plan, operators are likely to increase the speed of mobile Internet to 10gbps or higher, about 100 times the speed of the existing network, and faster than the existing fixed network. High-definition movies can be downloaded in seconds, even while walking. This will certainly lead to the further growth of mobile video business. In the 4G era, the scale and proportion of mobile network video users are growing rapidly. by the end of 2016, the number of mobile network video users is close to 500 million, and mobile network video users account for 91.8% of mobile Internet users. With the arrival of 5G to improve the speed of the network, the support of hardware is also improving. In conjunction with 3D imaging VR/AR, network video is expected to achieve more games in the future.

From the perspective of market demand, the current demand for artificial intelligence chips can be divided into three categories: first, the training market for major artificial intelligence enterprises and laboratory research and development; second, data center inference that mainstream artificial intelligence applications, such as Amazon.Com Inc Alexa, need to provide services through the cloud, that is, the inference link is placed in the cloud rather than user equipment. The third is to infer the market for smart phones, intelligent security cameras, robots / drones, autopilot, VR and other devices, and the device side infers that the market needs highly customized, low-power artificial intelligence chip products. For example, Huawei installed Cambrian IP in Mate 10's Kirin 970, which aims to achieve strong deep learning local computing power for mobile phones, thus supporting artificial intelligence applications that used to require cloud computing.

From the classification of chips, artificial intelligence chips through the attempt of CPU, GPU, FPGA, Cambrian artificial intelligence ASIC chips have come out, which is a major turning point in the development of artificial intelligence, the chip from the bottom of the hardware provides power for the development of artificial intelligence, and its effect far exceeds that of advanced algorithms. However, the development of artificial intelligence ASIC is still in its early stage. The fundamental reason is that once the ASIC is designed and manufactured, the circuit is fixed and can only be fine-tuned and cannot be greatly changed. The cost of R & D, design and production of hardware is very high, if it is not clear whether the application scenario is the real market, it is difficult for enterprises to try it rashly. In addition, the companies that can design chips suitable for artificial intelligence must have both artificial intelligence algorithms and be good at chip research and development, with higher barriers to entry.

The ideal artificial intelligence chip needs to have:

1. Programmability: adapt to the evolution of algorithms and the diversity of applications; 2. Dynamic variability of architecture: adapt to different algorithms to achieve efficient computing; 3. Efficient architecture transformation capability: 25f/s. 6. Low cost: access to home appliances and consumer electronics; 7. Small size: can be loaded on mobile devices; 8 easy to develop applications: no knowledge of chip recognition is required.

According to the above conditions, the current practices of CPU, CPU+GPU, CPU+FPGA and CPU+ASIC are not ideal architectures.

According to the research results of the Institute of Microelectronics of Tsinghua University, reconfigurable computing chips are the most suitable for artificial intelligence. Reconfigurable chips support software and hardware programming, and the hardware changes with the software, that is, software-defined chips. The scale of the software can be arbitrarily large, but the scale of the hardware is always limited, so (1) the software needs to be divided into modules adapted to the scale of the hardware, and run block by block. (2) the hardware structure must be dynamically variable to adapt to the function of each software module.

Therefore, if we compare reconfigurable computing with classical computing, the classical structure is rigid, while reconfigurable computing is functional and flexible. It focuses on ASIC in performance and processor in flexibility. It's both.

Although the reconfigurable chip is in the concept stage, there is still some distance from the real landing, but this provides a perspective for us to study the artificial intelligence chip, which may become the turning point of artificial intelligence.

6 computers: cloud computing and cloud services are the main theme of 2018

From the cost advantage point of view, the per capita GDP of China is already 1x6 of the per capita GDP of the United States, while ten years ago, the figure was 1max 20. If the per capita GDP can be regarded as a substitute for per capita income, then the labor cost of China's software industry has increased significantly, so that the industry has to think about the use of cloud computing, SAAS and other models to greatly improve labor productivity and reduce head-based project expenditure.

In addition, the per capita GDP of the United States has reached the level of US $10,000, and the proportion of the tertiary industry has reached 55-60% (at the end of the 1960s, about the level of China at present). The revolution of information technology (computer, Internet, cloud computing) continues to break the tradition and improve productivity, creating a number of software companies such as Microsoft Corp, Oracle Corp, Amazon.Com Inc, Salesforce.com Inc and other software companies for decades. According to McKinsey, the top 12 disruptive technologies that determine the economy in 2025 are not only the mobile Internet, but also the automation of knowledge work (which is the most apt description of SAAS), the Internet of things, cloud computing, advanced robots, automatic cars, genomes, energy storage technology, 3D printing, advanced oil and gas exploration, advanced materials and renewable energy. As we can see, SAAS and cloud computing almost occupy a very high position on the list.

In 2017, the domestic SAAS market was US $600 million and is expected to grow at an annualized rate of 25% in the future (global CAGR is 17%). In the public cloud market, China's market space will reach US $3 billion in 2017, of which the main part will be IAAS (represented by BABA), followed by SAAS and PAAS of US $600 million each.

Cloud computing has become a major growth driver in the data center industry. Cloud computing uses a network of remote servers hosted on the Internet (rather than a local server or personal computer) to store, manage, and process data. Today, almost all consumer data applications are delivered as cloud services, and enterprise application data is rapidly moving to cloud computing. By 2019, more than 80% of the workload will be handled by cloud data centers, while the rest will be handled by traditional data centers. Cloud computing requires a large number of physical sensor connectors, routers, power, signals and networks, which requires a large number of optical sensors and connectors. Therefore, we believe that, in addition to SAAS, software companies, the upstream of cloud computing, optical devices, optical chips, connectors and so on will directly benefit from the investment in the data center.

China Intermediate Xuchuang, Optical Technology, Hong Kong stock Hongteng Precision (6088.HK, the acquisition of avago's optical module part), 0877.HK Technology Group is the main player in this big market. With the launch of domestic 100G chips, China's optical chips and optical devices will enhance their competitiveness in the world, and are expected to extend their capabilities to industrial machine vision, automotive ADAS and other fields.

7 Internet: Battle Royale Games is expected to become a "LOL" product in 2018

On December 1, Tencent officially announced that he would act as the world's popular "PUBG" this year, and announced plans such as the exclusive operation plan and ecological co-construction strategy of "PUBG". "PUBG" is a PLAYERUNKNOWN'S BATTLEGROUNDS game jointly developed by Playerun known, the producer of games such as Bluehole and H1Z1. Each game has up to 100 players, and you can choose three modes: single player, double player and four-player team. At the beginning of the game, all players will be airdropped to a Jedi island, where there will be a shrinking security zone from time to time, and there will be poison gas outside the zone that will cause the player's physical strength to drop in the gas circle until he is "killed".

Therefore, all players need to enter the safe area as much as possible before being attacked by the gas, and have a showdown with other players they encounter along the way. In the end, only one person / team will win the game, and the word "Winner winner, chicken dinner" will appear on the winner's screen.

Since its launch in March 2017, PUBG has sold more than 22.8 million copies worldwide, topping steam sales for more than 38 weeks in a row. At present, "PUBG" has more than 2 million online players at the same time, far higher than the subsequent dota2 and csgo. At present, domestic players, who account for more than 1/3 of the total number of gamers, are willing to put up with high latency and unstable servers and pay an accelerator fee of 30 to 60 yuan per month to play the game, which sells for 98 yuan. This shows the popularity of PUBG. We believe that with the opening of the national service, the vast majority of more than 6 million domestic players will return, and more players who are still waiting because of network, equipment, servers and other problems will also join. "PUBG" has a huge market space.

The unprecedented game experience is expected to support the next popular game type after moba. "PLAYERUNKNOWN'S BATTLEGROUNDS" game is an open world system, its model can bring players a very rich game experience, we briefly summarize as follows:

The unknown inspires a strong sense of entry and curiosity.The basic rules of "PUBG" are simple and clear-the safe zone will certainly be reduced and try to survive longer; but the randomness of setting (random route, landing place, picking up objects, location of safe zone) and the randomness of participating players' behavior bring great contingency and unpredictability to the game, which combines "fighting with people and fighting with heaven", which greatly enhances the players' sense of substitution and curiosity.

Players can gain a high sense of achievement and excitement.There are as many as 100 players involved in each game, and if you can successfully "eat chicken", it means that you beat your opponent in 99-98-96, and the daunting challenge can inspire players' concentration and desire for survival. The process of the game has to go through many links, such as nothing-picking up equipment, escaping from the poison circle, player confrontation and so on. During the period of forbearance, killing and drug escape, players will be given a rich game experience and increased expectations for rewards after hard work. With the deepening of the game process, the excitement of players will continue to increase, and there will even be a surge of adrenaline in the final stage, and the sense of excitement and achievement after winning will be higher than that of moba and FPS games in most scenarios.

The entry threshold is low, and it is easier to attract new players.Traditional competitive games, such as League of Legends, dota2, csgo, etc., all know the number of teammates and opponents, fight on familiar maps and lines, and evolve more fixed strategies and higher entry barriers over a long period of time. Through diversified strategic choices and great randomness, PUBG weakens the importance of shooting skills in the game, lowers the entry threshold for beginners, and is easier to attract new players and even female players.

The game mode is rich and the social attribute is obvious.PUBG can choose single-player mode, double-player mode and four-player mode to meet the needs of different players. In the game, one-click in-team voice helps friends gang up and strangers make friends through the game. The popularity of the word "chicken eating" also reflects the obvious social attribute of the game.

Cracking down on plug-ins and building ecology will be the focus of Tencent's agent. A large number of data operations in shooting games need to be carried out on the local computer, so the plug-in is difficult to be completely eliminated. The popularity of "PUBG" has also led to the emergence of a large number of plug-in producers and users in China, seriously damaging the game experience of ordinary players and causing the number of bad reviews on games on steam to soar since October. Tencent has rich experience in the operation of shooting games and advanced big data's retrieval, analysis and processing capabilities. after acting National Service, whether he can step up efforts to crack down on plug-ins and create a good game environment for players, and around games, live streaming, events and other plates to create a complete ecosystem, will determine the life of this game to a large extent.

The profit of PC games refers to the existing model of steam.We believe that after Tencent acts as an agent for national service, "PUBG" is still likely to use its profit model of selling games + jewelry out of the box + transaction commission on steam. At present, the price of the game is 98 yuan per set and the key is 15 yuan per set. Steam will take 15% of the transactions on the market. At present, "PUBG" clothing and boxes are popular in the steam market, and if Tencent can launch characteristic accessories according to Chinese players' preferences in the future, it is also expected to continue this popularity.

The curtain of "eating chicken" on the mobile end has opened, and the number of users is growing rapidly. Recently, NetEase, Inc, Tencent, XIAOMI and other companies have launched a number of mobile games, the number of players has increased rapidly, of which the most popular NetEase, Inc "Wild Action" registered users has exceeded 100 million, and this DAU has exceeded 20 million, setting the fastest record for domestic mobile games with more than 10 million DAU. The rapid growth in the number of mobile chicken-eating mobile games and the scale of players reflects that chicken-eating games are also very attractive to mobile game players who pay attention to leisure and fragmented time.

At present, NetEase, Inc has taken the lead in the competition of chicken-eating mobile games with the two games of "Wild Action" and "Terminator 2: judgment Day". Tencent will soon launch two genuine "PUBG" mobile games in cooperation with Blue Hole: "PUBG: stimulating the Battlefield" and "PUBG: the whole Army attack". Coupled with the previous adaptation of "Desert Island Special training" and less-than-expected "Glorious Mission" based on CF mobile games, Tencent's hands of chicken eating mobile games will reach four. We think that "authentic" and "PUBG" mobile games will be Tencent's biggest selling point. With the accumulated popularity of "PUBG" and Tencent's leading position on the social platform, if its two games can achieve the goals of high quality and high degree of reduction, they are expected to achieve late attack and recreate the popular mobile games at the "Arena of Valor" level.

List of 8 parts of companies

The translation is provided by third-party software.

The above content is for informational or educational purposes only and does not constitute any investment advice related to Futu. Although we strive to ensure the truthfulness, accuracy, and originality of all such content, we cannot guarantee it.

2018港股投资策略报告：慢牛徐行还是前扬后抑？

2018 Hong Kong Stock Investment Strategy Report: Slowly Bull or Back Down?

1 美联储升息中的股票市场表现

2 国内：预计2018年GDP增速略降，国债收益率将呈现逐步下降态势

3 海外：基数效应的压力将于2018年逐步显现

4 2018年投资建议：看好金融、 原油、 电信、医药，创新关注云计算、 3D 结构光

1 the performance of the stock market during the Fed's interest rate hike

2 domestic: the growth rate of GDP is expected to slow slightly in 2018, and the yield of treasury bonds will show a trend of gradual decline.

3 overseas: the pressure of base effect will gradually emerge in 2018.

4 2018 investment advice: optimistic about finance, crude oil, telecommunications, medicine, innovation focus on cloud computing, 3D structured light

Risk Disclaimer

Statement

4 2018年投资建议：看好金融、原油、电信、医药，创新关注云计算、 3D 结构光