Last Night’s Markets
Buy The Dips (Carefully)
Some Other Views
Super Fund Lending
A Constant Investor Explainer: Afiniti vs Flamingo in the Battle for AI Bots
Research and Diversions
Last night the Wall Street indexes fell again – for the past three weeks the S&P 500 has had one rising day per week, and briefly last night it was in official correction territory, down 10%, before recovering a bit in the afternoon.
Today’s fall came after better than expected US GDP growth in the September quarter – 3.5% versus a consensus forecast of 3.3% – and was attributed mainly to Amazon’s and Google’s quarterly results falling short.
Of course, strong economic growth is mostly a downer for the market at the moment because it confirms that a December rate hike will happen and there’ll more next year, which is the main reason for the correction.
The third-quarter growth rate was led by consumption, which accounts for over two-thirds of the economy. Consumer spending accelerated to a 4% year-on-year increase during the quarter, up from the 3.8% rate notched up in the second quarter and higher than expected.
It’s pretty clear we are in the end stages of a nine-year bull market. Whether there might be another leg higher before it finally ends, or whether September was it, is hard to know as always.
It doesn’t really matter in my view: if a bull market is obviously close to ending, best to act as if it has ended. Trying to wring the last 10% out of it is picking up pennies in front of a fast-moving steamroller.
But just because a bull market ends doesn’t mean we are heading for a crash and/or the sort of brutal bear market of the 1970s. Of course, nothing is definite when it comes to predictions and no one has a perfect record (especially me!). The best anyone can come up with is an “on balance” view, and hope you’re roughly correct, and on balance I reckon markets will go sideways for a while.
That’s what I mean by today’s headline: sideways markets are for selectively buying the dips and carefully buying good companies cheap. This is not a cop-out, I hasten to add, with me sitting on the fence between optimism and gloom. Equity markets flat-line all the time, sometimes for years, as this 20-year chart of the All Ordinaries Accumulation Index shows – from October 1998 to now:
So, to sum up views I’ve been laying out here for a few weeks:
- The bull market is probably over because bond yields are rising and the because Fed and ECB are withdrawing liquidity and, in the Fed’s case, raising interest rates;
- Although a calamity can’t be ruled out, mainly because Donald Trump is US President, the economy and earnings look reasonably strong (in both the US and Australia), but not too strong to cause a very rapid rise in interest rates and therefore a steep decline in share values.
The stand-off between the US and China is a wild card and will get worse. As I’ve argued here, it’s an attempt by the Trump Administration to force US companies to shift their supply chains away from China and bestow the largesse of American consumers on other countries, to contain China’s rise.
But China can take a lot of the adjustment through a lower renminbi-US dollar exchange rate, which has already fallen a lot and could fall a lot more.
In any case, manufacturers will diversify their supply chains into other countries that haven’t had tariffs imposed, which would be bullish for them and not bearish for the US.
There are many reasons to think the investment environment is going to gradually deteriorate from here. And there are also plenty of other risks, like a disorderly Brexit, the brewing Italian crisis, or a breakdown in US-Saudi relations.
But US corporate earnings are still strong and interest rates remain fairly low (the real cash rate is still zero), the Fed is moving gradually and the bond market is showing no signs getting disorderly.
As for Australia, we don’t have much of an overpriced tech sector, so the spikes and dips here are both muted, and while a lot of the globally-facing industrials and miners look expensive, as I’ve argued here before, the banks are coming into value.
The Australian economy is chugging along OK, neither booming nor busting, employment growth is moderate, wages growth weak, and the cash rate is stuck where it is for another 12 months at least.
In some ways, Australia is the best place to be, and in that context I note that The Economist has a cover story this week headed: “Aussie rules: what Australia can teach the world.”
Here’s what the promotion for the story says: “Our cover this week holds out Australia as an example for the world. Rising incomes, low public debt, an affordable welfare state, popular support for mass immigration and a broad political consensus: in most of the rich world such blessings are a distant dream, in Australia they are a proud reality. The country has been growing for 27 years without a recession. The public finances are in excellent long-term shape. Half of Australians are either immigrants or the children of immigrants. What is Australia’s secret?”
As I remarked to Percy Allan: Goodness! They obviously haven’t been listening to Parliamentary Question Time.
The key risk, of course, is the property market. With clearance rates now well below 50% and a lot of investors over-extended, there is clearly the possibility of panic selling and bigger price declines than the 10-12% most analysts are predicting.
That would flow through to the banks and other lenders, as well as to consumer spending, and needs to be watched carefully. At some point the RBA might have to come to the rescue with a rate cut. Would it? Of course it would, although as usual it would be too late.
To conclude, I don’t think it’s time to head for the hills but to look for long-term bargains.
If you’re playing the market as a whole, good luck. It’s probably going to be choppy for a while.
My views on the market are well known to regular readers of the Overview, and are restated above after a tough week.
But I think it’s worth hearing from a few others I respect as well. These are all taken from notes or articles this week.
The investment year has been dominated by a widening gap between US and the rest of the world in terms of economic growth, earnings growth, equity performance, interest rates and, to a lesser extent, the dollar. What’s the chance that next year is all about these gaps reversing?
A gap has opened up between the US and the rest of the world on a number of fronts. The most important – because it drives the rest – is the diverging growth cycles. Most global cyclical indicators have rolled over in 2018. In the US, however, many cyclical indicators have made new highs – in some cases, all-time highs.
There has been a similar divergence in profit growth. To be fair, the better US EPS growth is partly due to the corporate tax cut. However, the most important factor has been the gap in sales growth: sales are up 8½% for US firms versus 2¾% for the rest of the world.
The earnings gap is not simply a matter of superior top-line growth: forecast US margins continue to rise, and are now at cycle highs. Buy-backs have not been a factor: buy-backs do not affect reported or forecast margins, nor do they affect index-level EPS series.
Put simply, many of the big trades this year have been driven by the widening gap between growth in the US and the rest of the world. The gap has widened in part because US growth has strengthened, and in part because growth elsewhere (notably Europe and China) has slowed. The forward looking point is the growth divergence is creating a major policy divergence that should in time lead to growth convergence. This in turn could lead to several of the market divergences narrowing. The most likely prospect is that US growth slows due to policy tightening. If, as many now expect (myself included), that policy tightening threatens a hard landing, then the equity convergence would likely reflect US returns falling to match the poor returns seen elsewhere.
Chris Weston of Pepperstone
Those … who like trading from the short-side will be looking at the deterioration in order book dynamics and seeing few buyers, amid a wave of selling and a capital exodus out of credit and equities and into cash, US Treasuries, German bunds, the JPY, and to a lesser extent gold.
The short volatility (vol) trade has been unwound at a sharp pace, and this has thrown additional weight to the notion that the ‘buy the dip’ mentality, held over the past two years, is over and we have entered a new paradigm. The lack of any counter-rally in the ASX 200 backs the idea that traders see heightened risks that we could see stocks gap lower tomorrow – asset price deflation if you will. (AK – that was Thursday. It didn’t gap lower on Friday; maybe next week).
The question of what is behind the moves leads me to think the elevated vol trade has further to play out, as there seems no clear circuit breaker.
The idea of a ‘Fed put’ has been a key talking point and there is merit to this as we have relied on the Fed backing off from rate hikes in 2015 and 2016 after a pullback in risk sentiment. The consensus seems to be this is closer to 2400 (for the S&P 500), which implies a far deeper correction in risk assets. However, the interest rate market is screaming out here, and we have already seen a full rate hike priced out of the Eurodollar futures curve between December 2018 to December 2020. The market is way ahead of the Fed here and looking at the rampant tightening of financial conditions, from higher Libor, implied vol, real yields against the drawdown in the S&P 500, and saying why on earth do you need to raise in this environment? This makes some sense and unless the core of the Fed start to acknowledge the risks to the real economy from tighter financial conditions, then we will get to a point where the rates market price in a 60% chance of a hike in December and no more hikes through to 2021!
Last week it was concern around moves in ‘real’ (or inflation-adjusted) yields, amid lower liquidity and higher funding costs that caused the anxiety. Well, all those issues are still there, they haven’t gone away, but they have been joined by calls for ‘peak earnings’, given the Q3 numbers have been buoyed by tax cuts. And it seems a genuine worry that the Fed is going to keep tightening policy when there are clear cracks viable in concentrated economic data points – such as housing.
We know the US housing market has been an issue for a while, it is reflected in lumber futures which are 53% lower since May and are coping a beating. We have seen the deterioration in the affordability dynamics for some time, just as we have in Australia and Canada. However, now we see weakness in existing and new home sales, and we can see the data flow and the various surprise indices (produced by Bloomberg) that measure the housing and the US business cycle ticking lower, which is tough when the Fed still have their foot on the tightening gas.
Europe is a worry. Its banking sector is in free fall, and the number of banks requesting cash (liquidity) from the ECB in its weekly repo operation known as the “MRO” has increased from 25 banks to 45 at the last count, taking the capital requests up from €2.7b to €7.7b. Weakness in the PMI figures have only heaped the misery on, but it is the banks that are the greater concern and this will only get worse should we see the Italian bond (BTP) spread widen against the German bund as traders continue to speculate that Italian banks will have to be the buyers of Italian sovereign issuance when the ECB stops buying their debt in December.
Shane Oliver, AMP
Let’s start with price to earnings ratios using historic earnings, ie, earnings reported for the last 12 months. In the US, this PE is currently around 21 times which is consistent with subdued medium-term returns going by the second scatter plot on the previous page. But in Australia at around 15.3 times it’s consistent with reasonable medium term returns according to the third scatter plot on the previous page. Of course, current levels for historic PEs in the US and Australia have both been associated with a very wide range in terms of subsequent returns historically. Furthermore, PEs based on historic earnings are not totally reliable given the cyclical volatility in reported earnings.
Given this, calculating the price to earnings ratio using 12 month ahead projected earnings are arguably more useful. These are shown in the next chart for global shares, the US and Australia. None are way out of line with their averages since the early 1990s but in a relative sense US shares on a forward PE of 16.1 times remain a bit expensive albeit less so than earlier this year, and global shares are a bit cheap trading on a forward PE of 14.2 times thanks to cheap markets outside the US.
The overall impression is that measured against their own history developed country share markets are not dirt cheap, but they haven’t been for several years now and they are not at overvalued extremes. The main risk relates to the US share market, but other markets’ valuations are reasonable. So while the pull back in shares we have seen over the last few weeks could go further yet – as worries around the US interest rates, US/China trade, rising oil prices, problems in the emerging world, President Trump and the US mid-term elections and the Italian budget remain – at least most share markets are not trading at overvalued extremes which would potentially accentuate downside risks.
Will Denyer, GaveKal
“Things are fine now, but they are going to get worse.” This is what I hear from commentators on US growth, from corporate managers talking about profit margins, and from Chinese exporters discussing the impact of the trade war. The same could be said of US financial conditions—they are fine now, but as interest rates rise they will deteriorate.
This is the big worry of investors in the once resilient US equity market. After yesterday’s -3% fall, the S&P 500 is down -9.2% since October 3, nearly matching its 1Q18 drawdown of -10%. Given that the US bull market is just four months shy of its 10th birthday, there are reasons to worry about its demise. Yet, while recognizing that we are closer to end of the cycle than the start, I do not think it is about to expire, and hence my advice to buy the dip.
In short, the US investment environment is getting worse, but only on the margin. It is still conducive to growth and rising asset values. In other words, the situation has gone from great to good. And I’m not sure that the situation will go from good to bad as quickly as this month’s market move suggests.
While multiple factors have contributed to the current negative sentiment, to my mind the key factor is the direction of US interest rates. The US equity sell-off started on October 4, the day after 10-year treasury yields jumped 10bp to a seven-year high of 3.15%. Similarly, February’s sell-off followed a substantial rise in US yields in the prior month and a strong employment report that bolstered expectations for short rate hikes. I continue to think that both short and long US rates will keep rising, and so I’m avoiding bonds and the most interest-rate sensitive parts of the equity market like autos, banks and property developers (see Equities In The Late Cycle, Revisited).
And yet looking at the overall US market, it should be noted that both short and long rates are well below my best estimate of the economy-wide return-on-invested capital. This means that financial conditions are still conducive to economic growth—and in that situation, equities tend to keep rising.
Ambrose Evans-Pritchard, The Telegraph
President Trump is right to identify the Fed as the cause of Wednesday’s Wall Street rout. He told a group of rich fundraisers in August that he had appointed Mr Powell on the mistaken assumption that he was a “cheap money” man.
Now he has discovered – like others before, Carter, Bush senior – that the institution is his nemesis. “The Fed has gone crazy. They’re so tight,” he said -after the Dow dropped 800 points on Wednesday night.
The Fed has been forced to hit the brakes to offset the inflationary fiscal blitz by the White House and Congress, a package of tax cuts and pork-barrel spending at the top of cycle when there is no slack in the economy.
Capital investment is surging at a 10 per cent rate. The US Conference Board’s indicator of consumer confidence is a whisker away from a peak last seen in 1968. The unemployment rate is the lowest in half a century.
Mr Trump is pouring petrol on the fire. Federal infrastructure -spending is flowing through when it is least needed. Fiscal loosening will push the budget deficit to 5 per cent of GDP next year (IMF data), four percentage points too high for this stage of economic cycle.
British economist Tim Congdon calls it a policy of “fiscal vandalism”. The output gap closed long ago. The stimulus feeds overheating. This risks a sudden catapult effect akin to the mid-Sixties when the coiled springs of the Great Inflation abruptly broke loose. On that occasion the Fed was caught napping.
Fed tightening is what ended every cycle since the First World War. Often the Fed is caught out at the inflection point and over-tightens. This may happen again. The growth rate of the real M1 money supply in the US has dropped to 1pc despite the economic boom. This leading indicator points to an economic slowdown next year.
Marcus Tuck, Mason Stevens
Although most people agree that the global equity cycle is in its late stage, opinions vary as to whether the end is nigh now or later. Whilst it is true that most cycles end with a monetary tightening induced recession, the trick is in getting the timing right.
The chart below shows the seasonally adjusted US civilian unemployment rate, which is currently 3.7% (as at September 2018). The shaded bars on the chart indicate US recessions. When the unemployment rate falls below 4% (the red line), recessions usually follow, but the lag can be 2 to 3 years later.
The normal transmission mechanism works something like this – a tight labour market causes wage growth to accelerate, which causes inflation to rise, which in turn causes the Federal Reserve to raise interest rates too much. How much is too much? That is only really known after the event, but a good guide to whether the Fed has gone too far is an inverted yield curve and high real interest rates (neither of which are present at the moment).
Wage growth is still fairly muted in the US, with average hourly earnings growing 2.8% year-on-year in September. Although the US unemployment rate is below 4%, the US underemployment rate stands at 7.5%. That has come down a long way from the 17% peak in 2009/10, but indicates that there is still some spare capacity left in the labour market.
Structural forces such as globalization and technological innovation continue to restrain wage and price growth in nearly all major economies, despite Trump’s efforts to the contrary. His tariff increases are a source of cost pressure, but businesses are adept at finding alternative suppliers.
Real interest rates are still quite low. With the Fed Funds rate currently 2% to 2.25%, and core US inflation at 2.2%, real short-term interest rates are effectively zero. The real 10-year US Treasury yield is currently around 1%, still well below the long-term average (see chart below). Even allowing for the possibility of a structural reduction in the equilibrium rate, the Fed will continue on its path of normalizing interest rates gradually, lifting real rates in the process.
Interest rates will continue to normalize but they are not overly restrictive at the moment. The end for the global equity cycle is definitely coming but may not be nigh.
Macquarie Wealth Management
- We are buying the dip in Australian equities. We don’t think the sell-off has been accompanied by a commensurate deterioration in fundamentals and as a result, the market looks oversold.
- This remains a tactical call. We think the emergence of downside risks (in particular rising interest rates) will keep the market rangebound in the months ahead.
- Our preference remains towards high-quality growth names which have been caught up in the sell-off. Value has failed to protect on the downside and many stocks in this space lack an earnings catalyst.
- We keep our overall equities allocation unchanged by raising Australian equities to overweight and cutting our European equities allocation to underweight on the back of a further deterioration in leading growth indicators.
Perhaps the starting point for a discussion about bargains should be AMP, after it fell an amazing 25% on Thursday, and has now halved in six months.
In preview: just because a stock falls a lot doesn’t make it a bargain, and while most analysts upgraded AMP to a buy yesterday, simply because it’s now well below their target prices and therefore looks oversold, I think AMP’s future still looks very challenging.
Specifically, there is a material possibility in my view that AMP disappears entirely; I simply can’t see the point of the company any more.
The $1.5 billion outflow from its wealth management operations in the September is likely to be just the beginning. The company’s image has suffered stunning damage which will take years to overcome, and in the meantime, the royal commission final report comes out in February and industry funds will continue to outperform – especially in a flat or falling sharemarket, since they own more unlisted assets.
The divestment of Life & Mature and NZ Life reinsurance, along with the future IPO of NZ Wealth, are good steps towards simplifying AMP, but the complexity of the deal and break-up highlights the intricacies of the structure.
Apart from the September quarter outflow, the main negative surprises on Thursday were: $80-90m in lost wealth earnings, largely because the life businesses being sold subsidise distribution; the lack of a plan to attack costs; and the second half dividend cut of about 30%.
AMP becomes a simpler company, which is a positive for the incoming CEO, Franceso de Ferrari, but much more exposed to wealth management, which is an overwhelming negative, given the probable continuing outflows, the end of vertical integration of advice and wealth management, which the AMP business is built on, and the looming end of grandfathered commissions.
Perhaps the best place to look for bargains in the months ahead will be the resources sector.
Not that they have underperformed lately: After five years of poor returns between 2011 and 2016, resources have delivered fantastic returns since then, outperforming the rest of the market on a total return basis by more than 93%.
This upswing differed from previous cycles in three important ways: they haven’t spent much on new mines and kept supply under control, there has been a lot of focus on operating costs and dividends have been increased significantly.
Right now, according to UBS analysts, resources offer a near 10% “true yield” (not including franking) over the next year, with $19 billion in dividends and $11 in buybacks expected.
Valuations are still below average despite the big run-up since 2016, and based on current spot commodity prices, earnings are about to be upgraded.
China is likely to respond to the trade war with infrastructure stimulus, which is likely to boost demand for commodities, especially iron ore and coal.
In addition to that, the Australian dollar is now undervalued and likely to stay that way as US interest rates rise and ours stay where they are.
And most of all, resources stocks and the banks, that is income stocks, and should be viewed that way. I don’t think they’ll go back to being cash eaters in a hurry.
I was asked on Thursday whether I prefer BHP or Rio Tinto – the answer was, and is, BHP. It’s selling on a free cash flow yield (share price to cash flow after capex) of 11% and is about to disgorge $11 billion to shareholders.
And on the subject of resources as income stocks, here’s a useful table from UBS:
Anthony Pratt and Paul Keating, among others, began a campaign this week to get Australia’s super funds to lend to businesses, since the banks don’t really want to do it anymore, or at least not much of it, and private companies like Pratt’s can’t go to the equity market for capital.
The industry fund-owned IFM Investors lent $100 million for 10 years recently to Lindsay Fox and Max Beck for their Essendon Fields project, and has now approved $2.2 billion in lending to businesses, although for them that’s an allocation of less than 2 per cent, so far, so not even a big toe in the water, just a pinkie.
So I don’t think the banks are quaking in their boots just yet.
But NAB executive and former NSW Premier, Mike Baird, declared this week that the push for super funds to lend to businesses was at a “tipping point”, because the amount of money in super now equates to, and probably exceeds the $2.8 trillion in total lending on home loans, non-financial corporations, credit cards and personal loans.
Anthony Pratt is reported to have said that 45 per cent of all pension funds in the United States were loaned to businesses. In Australia, he said, it’s 1 per cent, so there’s clearly plenty of room for growth here.
Forty-five per cent of all super funds would be about $1.25 trillion. According to Reserve Bank statistics, business lending in this country totals a bit less than $1 trillion.
So there is enough money in super to completely drown the nation’s businesses in cash and destroy the banks’ business lending divisions, not that there is any danger of that happening for a while.
Super funds making business loans is a good idea, since long term business loans are hard to come by from banks at the best of times, and are a good match for long term superannuation liabilities, and probably a better return than equities from here as well.
But the problem is that such loans are not liquid – the money once lent is there for the duration – and super funds must, absurdly, meet any request for redemption within three days, so the great majority of their investments must be liquid.
Under current rules, the only way super funds could do significant business lending would be via tradeable securities, that is – bonds. But Australia doesn’t have much of a corporate bond market, and therein lies a story.
I blame Jonathan Binns Were, the founder of Australia’s capital markets.
JB Were arrived in Melbourne in 1839, aged 30, set up JB Were and Company as an importer and exporter, then started dealing in gold and in 1853 began dealing in shares. He soon became a stockbroker, and was elected as the first chairman of the Melbourne Stock Exchange.
He was also the first chairman of the Chamber of Commerce, president of the Bible Society, he was on the Hospital Committee, the Immigration Board and was a director of many, many companies. In other words, he basically ran the place, and the capital market that he established in this country was based on equities.
America’s equivalent of JB Were was J Pierpont Morgan, who was born in Hartford, Connecticut, two years after Were arrived in Australia from England. While Were was a trader by nature, Morgan was a banker, and the capital market he helped establish in the United States was based on bonds, not equities.
I’m not suggesting the reason America has a huge and thriving bond market and Australia does not is only to do with the different personalities of JB Were and JP Morgan, but I am suggesting it’s a factor. More to the point, perhaps, they were men who reflected their surroundings, and their times.
Melbourne in the 19th century was a gold rush town populated by punters; New York was a town of bankers whose currency was debt and credit risk, and by 1900 JP Morgan & Co was one of the most powerful banking houses in the world.
The development of Australia’s mining industry in the early 20th century in Broken Hill, Mt Morgan and Mt Isa reinforced the bias of Australia’s capital market towards equity, not debt; the only fixed interest securities being issued were preference shares and debentures. A bond market never got off the ground.
In the US, on the other hand, it was all about bonds, especially after the stock market collapse of 1929. In that year and through the years of the Great Depression, bonds kept producing returns above 3 per cent and became wildly popular.
And in the 1970s and 80s, when Australia was going mad for spec stocks and takeover artists like Alan Bond, America was enthralled by junk bonds – high-yield bonds issued by non-investment grade companies, made popular by Drexel Burnham Lambert and Michael Milken.
In the US, even the speculation was done with bonds rather than equities – until the internet arrived in the 1990s and the speculation switched to equities.
These days the US bond market totals more than $30 trillion and total pension fund assets are about $20 trillion. If Anthony Pratt is right, that means about $9 trillion, or 30 per cent of the bonds are owned by pension funds, which sounds high but could be right.
Australia has a long way to go, and a lot of history to overcome, to come close to equalling that sort of pension fund allocation to corporate debt.
And it’s really up to the super funds themselves and, in particular, the industry funds that are now carrying all before them in the aftermath of the royal commission.
Not only do they have the money now pouring in from the discredited bank-owned super funds, but they also have the business lending market opportunity as the banks abandon all forms of risky loans.
Australia’s industry super funds have created a large and sophisticated market for infrastructure investments and consistent returns of 10 per cent. Now they should turn their minds to creating a corporate bond market for the same reason.
Prof Cara MacNish
Alan, subscriber Jamie asked a great question on the Facebook Livestream:
Absolutely superb interview with Zia of Afiniti, and his interpretation of what’s real AI….I’m embarrassed that I too have been caught up in the hype of AI and machine learning, algorithmic processing versus sentient/semantic processing.
With that in mind, do you feel that Catriona Wallace, CEO of Flamingo AI, misled TCI listeners? In your interview, Dr Catriona mentioned numerous times about the ability of FGO to be able to use AI/machine learning to deal with customers when really their system is just programmed questions and answers, or pattern recognition, not true AI/machine learning.
The differences between Afiniti and Flamingo highlight one of the hot debates in applied AI just now: Will AI replace human jobs, or will it help humans to work better?
In the blue corner, Catriona Wallace of Flamingo AI (May 24) is seeking to create fully automated virtual assistants or in her words “chatbots on steroids”, to guide customers through financial services product purchases and customer service.
In the red corner, Zia Chishti of Afiniti (October 16) claims that “the actual quality of human live voice recognition is terrible at best and there’s no danger to human replacement any time soon”. Instead, Afiniti seeks to improve the matching of customers to human call centre operators.
Before we weigh in on the battle, let’s answer Jamie’s specific question: did Wallace mislead TCI listeners? To answer this we need to take a brief look at two ways that machines learn classifications. In fact, they’re also two ways that humans learn.
The first is called supervised learning. Let’s say a preschooler from the burbs, Nina, is taken to a farm to learn about animals. The teacher shows Nina some working dogs. Nina sees some of the attributes they have in common, and forms a mental picture of the class ‘dog’. The same occurs when the teacher shows Nina some sheep.
Later Nina goes out for a wander on her own and sees many more examples of animals. When she sees a new animal, say a labradoodle, she endeavours to work out which of the animal classes she knows about that it best fits into – or which examples its closest to. Because she doesn’t have a teacher with her, she needs to try to refine her own classifications of the animals, perhaps regrouping the examples she’s seen for a better fit. We call this unsupervised learning.
The classification task is one of the most fundamental and important tasks in AI. If I see/hear/sense some ‘thing’, what kind of ‘thing’ do I think it is?
For Flamingo AI, the things might be customer queries. Flamingo does exactly what Nina did. First, they hire market research company Appen (the teacher) to train their bot Rosie by surveying customers for sample questions, and matching them with the right answers (supervised learning).
However, because language is so flexible, there are many different ways of asking similar questions. When a client asks a question that Rosie hasn’t been seen before, their bot’s job is to work out what class of questions it belongs to. As Rosie accumulates these new conversations, she is doing unsupervised learning.
The classification task is a fundamental AI, so Dr Wallace is not misleading TCI listeners when she says they are endeavouring to use AI.
In fact, as Zia admits in many ways it’s a harder task than what Afiniti are trying to do. The best chance of making
it work is by choosing very specific domains with high potential yield. Its no coincidence then that Flamingo have chosen to focus on financial products such as insurance.
Afiniti, on the other hand, are aiming at a broader market by keeping humans at the end of the line. To stretch the metaphor, rather than train up one kid to be an expert, Afiniti are taking a whole busload of kids and trying to direct each query to the kid that knows best. Their AI must learn which kids can answer which kinds of questions best (and perhaps most politely). A less ambitious use of automation, but more general purpose and robust.
So who will win in the near term? Well, it’s really a false dichotomy, egged up a little by zealous CEOs. Both can proceed at their own rate in their own domains.
In the longer term though? While AI augmenting human capacity is more palatable, it would take a brave punter to back Zia’s prediction that there is no danger of bots increasingly replacing humans.
– Thanks again Cara.
Here’s a pretty comprehensive list of all the reasons the market fell this week. The correct one is bound to be in there somewhere – then the author can say: See! I was right!
The Bombs addressed to Obama, Clinton, Soros, and the history of Anti-Soros hate-mongering.
Jamal Khashoggi: murder in the consulate. This is a good piece in The Guardian putting together what happened, directly fingering the Saudi crown prince.
The biggest takeaway from the Royal Commission hearings may be on governance. The impact on the mindset of boards and management towards risk aversion creates scope for disruption to have a bigger impact than otherwise. It’s possible that risk aversion will create a space, or wedge, in financial services that may be filled by disruptive firms.
An essay in Aoris Investment Management’s quarterly report: “The terms ‘growth’ and ‘value’ … suggest you can divide stocks and investors into two opposing tribes, like Democrats and Republicans or the Montagues and the Capulets. These distinctions are nonsense.”
Terrific white paper from InvestSmart (yes, the owner of Eureka Report): “How Fees Can Destroy Your Wealth – understanding the total cost of investing.”
John Mauldin: the real cost of low fee funds.
Why Trump supporters will believe any lie he tells – they are immune from the truth.
Anonymous, threatening letters are being sent to UK homes to try to stop activities that the Chinese government disapproves of.
The second wave of big grid batteries is about to come online – in Victoria.
The NBN won’t be finished on time. Simple as that.
Good piece by Tony Windsor about Scott Morrison: “Leadership is not something that comes with numbers and games, it comes from a genuine desire to do good for the community. The prime minister is yet to display this quality.”
5 stocks to buy on weakness (Tabcorp, Bega Cheese, Saracen, Atlas Arteria, Baidu).
The Roman ‘Brexit’: how life in Britain changed after 409AD. The Roman legacy crumbled rapidly. Industrial pottery manufacture vanished by about 420 AD. Villas, some of which achieved a peak of grandeur in the 4th century, were abandoned.
3D printed lithium-ion batteries are now a reality. “Imagine a world where lithium-powered consumer products can be created in any shape or size – as could the batteries that keep them running.”
Atomic Swaps, the ultimate guide. Do you need to know? Well, maybe not. “Atomic Swaps has the potential of completely revolutionizing the money transfer system in the crypto world. To put it in simple terms, atomic swaps will enable people to directly trade with one another wallet-to-wallet.”
Vancouver: The City That Had Too Much Money. “The money being frenetically shuffled by millions of wealthy Chinese into safe assets abroad, in defiance of their country’s capital controls. Since mid-2014, capital flight from China may have totalled as much as $800 billion, according to estimates from the Institute of International Finance. In Vancouver, the tidal wave has wrought a dramatic economic, demographic, and physical transformation.”
100 websites that shaped the internet as we know it. “What does a spot on this list mean? It certainly doesn’t mean “best.” A number of them are cesspools now and always have been. …we set out to rank the websites — not apps (like Instagram), not services (like PayPal) — that influenced the very nature of the internet, changed the world, stole ideas better than anyone, pioneered a genre, or were just really important to us.”
The silver lining in China’s slowdown. (Steel production is hitting new records).
The Archivists of Extinction. Laws that protect buildings of architectural or historic importance apply only to buildings more than 50 years old; they assume that nothing built recently is worth preserving. If future generations want to know how a McDonald’s or a K-Mark looked, they will have to work from photographs
ANZ chief Shayne Elliott: “As the realities of the digital economy unfold, it is clear banks can’t do everything themselves – we need to collaborate and partner with others in a way that makes sense for our customers, bringing in skills, technology, innovation.”
This is quite interesting. A digital consultancy named Cognizant has created something called the “Cognizant Jobs of the Future Index”, which measures the changes in demand for a set of 50 jobs of the future: 45 actual and 5 proxy jobs. It also includes eight additional sub-indices that represent families of similar jobs. The CJoF Index score represents the sum of the current quarterly total of U.S. job openings divided by the quarterly total of U.S. job openings in the 3rd quarter of 2016 (the base year for the Index).
“This is not the beginning, much less the beginning of the beginning. Computers have been around for the better part of a century, personal computers since the late 1970s. Looking back we see differences in degree, not in kind.”
“I once wrote a user manual for a handheld engine-diagnostics module for Daewoo cars. I know “hard to use” when I see it. iPhones are hard to use.”
John Milton was a snob and a narcissist; a kind of 17th-century prototype for the metropolitan liberal elite. “And yet I love him. He loathed tyranny and made the case for meritocratic replacement of authoritarian kings and magistrates. He believed that intellect, rigour, and reasonable debate could defeat evil. He stood up for the right of miserable married couples to divorce in pursuit of happiness – a principle we are only now enacting into law. And when critics silenced him through censorship, he fought back, writing one of history’s most compelling defences of free speech.”
So close, and yet so far: a brief history of science fiction’s love affair with Mars.
Amazing footage of humpback whales bubble feeding near the Antarctic.
11 toxic ideas you should get rid of to have a better life. I usually hate these sort of pieces, but this one’s not bad, I must say. No.1 toxic thought: You can’t change other people. How true.
Race doesn’t come into it: The more we learn about human genetics, the more complex its workings appear. “…part of your DNA is actually viral DNA from retroviruses that once upon a time inserted their DNA into human reproductive cells, thus becoming heritable. At the level of DNA, humans are actually a mash-up of different species”
On the internet we overshare about our personal lives and fail to understand those of others. Narcissism spreads; empathy vanishes.
Historian David Frye on the building of big walls. “The first border walls are found in the late 2000s BC in Mesopotamia, when there were two different lifestyles developing. The wallers are workers who build things, and identify themselves by their civilian occupations, and secure themselves by building structures that protect them even when they are sleeping at night. Outside the walls you have a very different sort of society, peoples inured to the dangers of living in an un-walled world, peoples we generally refer to historically as barbarians, like the Huns, the Goths, or the Mongols”.
A visit to Britain’s only medicinal leech farm. “The leeches I encounter are freshwater, bloodsucking, multi-segmented annelid worms with 10 stomachs, 32 brains, nine pairs of testicles, and several hundred teeth that leave a distinctive bite mark.
Banksy’s autoshredding stunt reinforces how contemporary art is not so much about art but the documentation of an event.
John Prine: American oracle.
How McLaren learned to treat its pit crew like Olympic athletes. “What separated the best athletes from the rest was the ability to zero in on the correct information even when under overwhelming pressure. In the case of the pit crew, focus should be on the nut and the wheel.”
It’s five years today since Lou Reed died, aged 71. Here he is doing Sweet Jane, live, including some very nice guitar work from someone else.
And today is also violin master Niccolo Paganini’s birthday (1782). Here’s his Caprice No.24 – it’s like watching an early version of a guitar solo by a combination of Joe Bonamassa and Eric Clapton.
Brexit protest, best placard:
If you missed #AskAlan on our Facebook group this week (or if you don’t have access to Facebook) you can catch up here. And we’ve just given the Facebook Livestream its own page where you can also opt to just listen to the questions and answers.
If you’re not on Facebook and would like to #AskAlan a question, please email it to email@example.com then keep an eye out for the Facebook Live video in next week’s overview.
By Craig James, Chief Economist, CommSec
Australia: All eyes on inflation indicators
- Quarterly inflation figures feature in Australia in the coming week. In addition there will be keen interest in the monthly and quarterly retail trade results.
- The week kicks off on Monday in Australia when CommSec releases its quarterly State of the States report – an economic performance report of state and territory economies.
- On Tuesday the regular weekly reading on consumer confidence is published by ANZ and Roy Morgan. And the Australian Bureau of Statistics (ABS) releases building approval data for September.
- Also on Tuesday, Reserve Bank Assistant Governor Michele Bullock, will speak at the Commonwealth Bank Global Markets Conference.
- On Wednesday, the ABS releases the quarterly Consumer Price Index – the main measure of inflation in the economy. At this stage we expect that prices rose by 0.3 per cent in the September quarter, causing the annual inflation rate to ease from 2.1 per cent to 1.8 per cent. The “underlying” annual inflation rate may print near 1.9 per cent. Petrol prices rose by 1.5 per cent in the quarter after a 6.9 per cent increase in the June quarter.
- Also on Wednesday the Reserve Bank releases the September private sector credit figures (effectively the value of outstanding loans).
- On Thursday a bevy of indicators are released. CoreLogic releases its latest data on home prices. Home prices may have fallen around 0.5 per cent in the month with Melbourne prices down 0.7 per cent.
- Also on Thursday the ABS releases data on international trade (exports and imports) as well as data on export and import prices. Also both AiGroup and CBA will issue latest survey results of purchasing managers in the manufacturing sector.
- Australia has posted trade surpluses in the past eight months and in 14 of the past 16 months.
- And on Friday the ABS will release the September quarter readings on business inflation (Producer Price Indexes) as well as retail trade data for both September and the September quarter.
Overseas: US employment report hogs the limelight
- In the US in the coming week the focus will be the latest US jobs report on Friday. But there will be indicators covering inflation, manufacturing activity, home prices and auto sales. The Chinese purchasing manager survey results are issued Tuesday and Wednesday.
- On Monday in the US, the September data on personal income and spending are released. But other features of the data are the inflation measures – especially the core personal consumption deflator (excludes food and energy prices). The core PCE deflator is the Federal Reserve’s preferred inflation measure.
- On Tuesday in the US the S&P/Case Shiller measure of home prices is released together with consumer confidence and the regular weekly chain store sales figures.
- On Wednesday in the US, ADP releases its October report on private sector employment, derived from payroll records. Analysts expect that payrolls rose by around 200,000 in the month.
- Also on Wednesday in the US, the quarterly employment cost index is released together with the regular weekly data on mortgage applications.
- On Thursday in the US, the ISM manufacturing index for October is released with data on construction spending, new auto sales and the usual weekly data on claims for unemployment insurance (forward indicator of the job market).
- On Friday in the US is the highlight of the week’s reports – the non-farm payrolls or employment report. And it is clear that analysts will scrutinise all the key components. Employment may have lifted by around 180,000, continuing the run of strong job gains. The unemployment rate may have held at a 48-year low of 3.7 per cent. And wages may have lifted 0.3 per cent to keep annual wage growth near 2.8 per cent.
- Also on Friday in the US is data on international trade and factory orders. Analysts expect that the trade deficit narrowed from US$53.2 billion to US452.4 billion in September.
- In China, the National Bureau of Statistics releases the “official” purchasing manager survey results for manufacturing and services on Wednesday. The private sector Caixin manufacturing survey data is on Thursday.
- Another big week of US earnings reports lays ahead. Companies expected to report earnings this week include:
- On Tuesday: Baker Hughes, Coca-Cola, General Electric, Pfizer, Ebay, Facebook.
- On Wednesday: General Motors, Yum! Brands, Tesla.
- On Thursday: DowDupont and Apple.
- On Friday: Chevron and Exxon Mobil.
By Shane Oliver, Head of Investment Strategy and Chief Economist, AMP Capital.
Investment markets and key developments over the past week
- Most major share markets fell over the past week as worries about the global outlook continue. Chinese shares rose though thanks to stimulus measures. Bond yields fell on safe haven demand and while iron ore prices rose, oil and metals fell. The $US rose and this weighed on the $A which fell to the lowest since early 2016.
- The share market correction continues with increasing concerns about global growth. From their recent highs to recent lows global shares have fallen about 9% and Australian shares nearly 11%. It’s still too early to say we have bottomed but we remain of the view that it’s not a major bear market. The worry list of rising US interest rates, the US/China conflict, a correction in tech stocks, problems in the emerging world, the US midterm elections and the Italian budget is continuing to drive shares down and as we have seen in the past this is morphing into another global growth scare with investors latching onto companies that had negative profit news in the US and declining Eurozone PMIs. Shares are technically oversold again and so may see a bounce, but a circa 20% fall as occurred through the 2015-16 global growth scare is possible and this would likely require some sort of global policy reaction to turnaround (eg the Fed hitting the pause button and the ECB extending QE – China is already easing).
- But the following key points are worth bearing in mind:
- Corrections are normal – global and Australian shares saw multiple pullbacks ranging from 7% to 20% since 2012.
- The main driver of whether we see a correction or even a mild bear market (say a 20% fall) as opposed a major bear market like the GFC is whether we see recession in the US and right now this still looks unlikely as we haven’t seen the sort of build-up in excess that precedes a US recession. Don’t forget that the share markets often overreact to risk and the old Paul Samuelson saying that “share markets predicted 9 of the last 5 recessions”.
- Selling shares after a big fall just turns a paper loss into a real loss.
- When shares fall in value they become cheaper and offer better return prospects so in this sense pull backs are good.
- While the value of shares has fallen dividends haven’t and so if its income you are after it hasn’t changed if you have a well diversified portfolio. In fact, the grossed up dividend yield on Australian shares is now around 6%.
- Shares bottom when everyone is in a panic. I don’t know when that will be but the trick is too look out for when the crowd gets very negative.
- Finally, to be a successful investor you need to keep your head and that gets hard it times like the present when negative news reaches fever pitch. So it’s best to turn down the noise and chill out a bit.
- These points are covered in more detail here.
- There is just one other thing worth mentioning. October is known for share market volatility. Sometimes its referred to as a bear killer given that in the US it often sees a share fall and then a rebound. But the point is that the share market traditionally strengthens through November and December in the US (and December in Australia). This is particularly the case in years of US midterm election (which is the second year of the US presidential term) particularly once the election uncertainty is out of the way. In fact the US share market hasn’t declined in the 12 months after a midterm election.
- Is this week’s The Economist cover – headed “Aussie rules” and showing a kangaroo bouncing through the clouds – the kiss of death for Australia? It reminded me of their 2009 cover showing Brazil/Christ the Redeemer blasting off only to fizzle back into the mountain in a 2013 cover. Worth checking the Magazine cover indicator.
Major global economic events and implications
- US economic data was a bit mixed. Home sales data remained weak consistent with the softness seen in other housing activity indicators lately but home prices are continuing to rise, durable goods orders were okay, business conditions PMIs rose in October and jobless claims remain ultra-low. Overall September quarter profit results have been good so far with 84% of results beating on earnings, 59% beating on revenue and earnings growth expectations moving up to 23% yoy. However, the level of earnings are down and investors have latched on to those who disappointed.
- The ECB made no changes to monetary policy with none expected and appeared to play down recent softer data. It remains on track to end QE in December, but it did refer to the possibility of using another round of cheap bank financing (LTROs if needed) and rate hikes still look to be a long way off. Meanwhile, Eurozone business conditions PMIs fell again in October albeit to a still reasonable 52.7 but adding to concerns that Eurozone growth is continuing to slow.
- The PMIs across the G3 are shown below. In short, the US is tracking sideways, Japan is possibly moving up and only Europe is seeing a downtrend. So no sign of a major developed country growth downturn – well at least not yet anyway!
- Details of personal tax cuts were announced in China and look like being bigger than expected at 0.5% of GDP.
Australian economic events and implications
- It was a quiet week on the data front in Australia, but what was released was soft. The CBA preliminary business conditions PMI fell to 51 for October down from 58 18 months ago and skilled vacancies fell for the sixth month in a row.
What to watch over the next week?
- In the US, the focus will be back on jobs with October payroll data to be released Friday expected to show solid jobs growth of 190,000, unemployment remaining at 3.7% and an in wages growth to 3.1% year on year. Meanwhile consumer data (Monday) is expected to show a solid rise in real consumer spending and the core private consumption deflator inflation remaining at 2% year on year, home prices are likely to show further gains and consumer confidence is likely to remain high (both Tuesday), September quarter growth in employment costs (Wednesday) is likely to remain at 2.7% year on year, the October ISM index (Thursday) is likely to remain strong at around 59 and the trade deficit (Friday) is likely to get a bit worse. September quarter earnings reports will continue to flow.
- Eurozone September quarter GDP data due Tuesday is expected to show moderate growth of around 0.3% quarter on quarter or 1.% year on year, unemployment is likely to have remained at 8.1% in September and core inflation for October is likely to have edged up to 1% yoy (both due Wednesday).
- Japanese jobs data (Tuesday) are likely to remain strong but industrial production (Wednesday) is expected to be soft.
- Both the Bank of Japan (Wednesday) and the Bank of England (Thursday) are expected to leave monetary policy on hold.
- Chinese PMIs for October (Wednesday & Thursday) will be watched for signs any further slowing in growth.
- In Australia the focus will be on September quarter consumer price inflation data (Wednesday) which is expected to show headline inflation of 0.4% quarter on quarter or 1.9% year on year with higher fuel prices and tobacco excise only partly offset by higher childcare rebates and lower electricity and gas prices. Underlying inflation is likely to remain subdued at 0.4%qoq or 1.9%yoy. Meanwhile, expect a 3% bounce back in building approvals (Tuesday) after their plunge in August, continued moderate credit growth (Wednesday), CoreLogic data for October (Thursday) to show a further decline in home prices, the trade surplus (also Thursday) to fall slightly and September retail sales (Friday) to show growth of 0.2%. Business conditions PMIs will also be released Thursday.
Outlook for markets
- We continue to see the trend in shares remaining up as global growth remains solid helping drive good earnings growth and monetary policy remains easy. However, the risk of a further short-term pull back is high given the threats around trade, emerging market contagion, ongoing Fed rate hikes and rising bond yields, the Mueller inquiry, the US mid-term elections and Italian budget negotiations. Property price weakness and election uncertainty add to the risks around Australian shares.
- Low but rising yields are likely to drive low returns from bonds, with Australian bonds outperforming global bonds.
- Unlisted commercial property and infrastructure are still likely to benefit from the search for yield, but it is waning.
- National capital city residential property prices are expected to slow further with Sydney and Melbourne property prices likely to fall another 15% or so, but Perth and Darwin property prices at or close to bottoming, and Hobart, Adelaide, Canberra and Brisbane seeing moderate gains.
- Cash and bank deposits are likely to continue to provide poor returns, with term deposit rates running around 2.2%.
- While the $A is now fallen close to our target of $US0.70 it likely still has more downside into the $US0.60s as the gap between the RBA’s cash rate and the US Fed Funds rate pushes further into negative territory as the US economy booms relative to Australia. Being short the $A remains a good hedge against things going wrong in the global economy.