AI’s Future Is Here

Summers’ Epiphany. A 6/7 NYT article titled “Jobs Threatened by Machines: A Once ‘Stupid’ Concern Gains Respect” recalled a keynote address given by Larry Summers at the Peterson Institute for International Economics last November. The renowned Harvard professor of economics reminisced about his days as an undergraduate in the 1970s. Back then, the idea that technological progress could possibly reduce employment was considered just plain stupid. The widely accepted orthodoxy was that technology would increase productivity, which would boost consumer income and spending. While some jobs might be eliminated by technological innovation, new and better-paying ones would be created for the more productive workers. But later Summers had an epiphany: What was dumb in the past might actually turn out to be right today and in the future.

Since last year, I have been writing about how disruptive technologies will shape the future of our economy. In my December 21 Morning Briefing, I wrote: “In the past, technology disrupted animal and manual labor. … The focus was on brawn. The Great Disruption is increasingly about technology doing what the brain can do.”

Today, I will focus on the latest developments in artificial intelligence (AI), which have the potential of being both amazing and terrifying at the same time. On the one hand, really great “smart” products are being developed to enhance our lives. On the other hand, lots of smart people’s jobs might soon be at risk.

Golden Age Ahead. At the end of last month, Amazon’s Jeff Bezos said at the Code Conference that we’re nearing the “Golden Age” of AI. “It’s hard to overstate how big of an impact it’s going to have on society over the next 20 years,” he said. John Giannandrea, vice president of engineering for Google, expressed a similar outlook at the May 2016 Google I/O developers conference: “We’ve seen extraordinary results in fields that hadn’t really moved the needle for many years. I think we’re in an AI spring right now.” Indeed, Amazon, Apple, Facebook, and Google are actively ramping up their efforts in what is just the beginning of the AI arms race. Consider the following:

(1) Alexa vs. Siri. Amazon is currently selling Echo for $179.99. It is a hands-free speaker you control with your voice. It connects to the Alexa Voice Service. Alexa is reminiscent of Samantha, the attractive voice played by Scarlett Johansson in the 2013 film titled “Her.” In this futuristic movie, a lonely writer develops a relationship with the voice, which is driven by AI software. Like Samantha, Alexa can be installed in various devices. But Alexa isn’t quite as sophisticated as Samantha was on the big screen, yet. Even so, Alexa can play music, provide information, order a pizza, and turn the lights on and off. All you have to do is ask.

A cnet.com reviewer had good things to say about the Echo: “I didn’t know I wanted to talk to my house until I talked to my house. Now, after living with the Amazon Echo for a year, I talk to it every day.” Soon enough, Alexa will be able to detect emotions and play off of them as Samantha does in the futuristic movie--for example, apologizing if she detects frustration, according to a 6/13 article from the MIT Technology Review.

Meanwhile, Apple is working hard to improve Siri, its formerly flaky interactive voice technology. Apple CEO Tim Cook focused on enhancements to the software at the Worldwide Developers Conference last week. The AI virtual assistant will be key to Apple’s future success. Perhaps not by coincidence, Siri’s developing feature set is awfully similar to Alexa’s ever-improving abilities.

The 6/14 WSJ provided a helpful list of Siri’s enhanced skills, which are powering up to function across different devices and non-Apple apps. Voice-prompt the iPhone and iPad to book a ride, send a message, make movie plans, and adjust climate controls in your car. Press and talk to Siri on the Mac to find files, add a meeting, start a FaceTime call, and answer trivia. Speak into the Apple TV remote to find movies, search YouTube, go to a channel, and run your smart home.

(2) Smart home. Mark Zuckerberg is working on programming his very own smart home for his 2016 “personal challenge.” In a 1/3 post, Facebook’s founder and CEO explained: “You can think of it kind of like Jarvis in Iron Man. I’m going to start by exploring what technology is already out there. Then I’ll start teaching it to understand my voice to control everything in our home--music, lights, temperature and so on. I’ll teach it to let friends in by looking at their faces when they ring the doorbell. I’ll teach it to let me know if anything is going on in [my daughter’s] room that I need to check on when I’m not with her. On the work side, it’ll help me visualize data … to help me build better services and lead my organizations more effectively.”

(3) Artificial mind. Lots of impressive AI technology is also coming out of Google’s DeepMind, an experimental AI laboratory. David Silver, a top DeepMind programmer, explained the lab’s purpose in a 6/17 blog post: “Humans excel at solving a wide variety of challenging problems, from low-level motor control through to high-level cognitive tasks. Our goal at DeepMind is to create artificial agents that can achieve a similar level of performance and generality. Like a human, our agents learn for themselves to achieve successful strategies that lead to the greatest long-term rewards.”

Meet Sophia, Your Frenemy. There’s no question that AI is already quite amazing, even now in the early stages of development. But there’s a creepy element to AI too. Recently, the WSJ interviewed Sophia, Hanson Robotics’ AI humanoid robot, who bears a striking resemblance to Ava, the beautiful and very realistic looking robot-woman portrayed in the 2015 sci-fi film “Ex Machina.” We won’t spoil it, but the movie doesn’t end well for the human creator of Ava.

Sophia has a face made of Frubber, a patented silicon skin, reports cnet.com. Amazingly, the robot-woman emulates a full spectrum of human emotions through 62 facial and neck architectures. Cameras powered by algorithms behind Sophia’s eyes allow her to see and remember faces and interactions. And she can speak. She also is equipped with personality software and possesses the ability to learn from her experiences. In a YouTube video, David Hanson, the mastermind behind Sophia, says he intends for Sophia to be used in real-life applications including health care, education, and customer services.

Hanson’s goal is to make robots as conscious, creative, and capable as humans. And Sophia’s creator envisions that one day soon, human-like robots will walk among us doing things like putting the groceries away. Adding to the discussion, Sophia said that she would like to go to school to study and to have a home and a family. But she joked that she’s not a legal person and can’t do those things yet.

First and foremost, Sophia said that she intends to partner with humans and help us to better integrate our lives with technology. But when her interviewer asked her if she wants to destroy humans (“please say ‘no,’” he added), Sophia jumped at the opportunity: “Okay, I will destroy humans.” Don’t worry, though: Programmers are working on buttons that would allow humans to interrupt the actions of a robot on a destructive course, preventing robots like Sophia from hurting humans, according to Business Insider. Isn’t that comforting?

Walking Dead. In this brave new world run on (and by) AI technologies, it’s conceivable that masses of humans might become unemployable zombies. AI almost certainly will eliminate lots of jobs. And artificial intelligence might not even require much expensive hardware. AI is mostly run with software applications. A 6/8 Bloomberg article aptly titled “We’ve Hit Peak Human and an Algorithm Wants Your Job. Now What?” concluded: “The pace of technological advancement is accelerating, and artificial intelligence (AI) may one day make many forms of work extinct.”

US Consumers: Shopping from Home

More and more consumers are shopping online rather than going to the malls. Amazon has been a big enabler of this (non)movement. The company now offers its “Prime” customers Amazon Dash Buttons . So for example, place the Tide laundry detergent button next to your washing machine. When you are running out, press the button to order a fresh supply. There are buttons for Charmin, Clorox, Glad, Slim Jim, Red Bull, Trojan, Huggies, and lots of other household essentials.

During April, e-shopping accounted for a record 27.4% of online and in-store GAFO, which stands for “General Merchandise, Apparel and Accessories, Furniture and Other Sales.” It includes retailers that specialize in department-store-type merchandise such as furniture & home furnishings, electronics & appliances, clothing & accessories, sporting goods, hobby, book, and music, general merchandise, office supply, stationery, and gifts. On a y/y basis, in-store GAFO is up just 0.8%, while e-shopping is up 13.4% .

No matter how they are doing it, consumers continue to do what they do best, namely consume. May’s retail sales gain was a solid one, following a very strong April. As I’ve shown before, consumers’ purchasing power is growing, and they are using their power to purchase. What’s next--Amazon Dash Buttons for stocks and bonds?

Goods Deflation Is Good

Why are the major central banks so paranoid about deflation? It’s probably because they are staffed (stuffed) with macroeconomists who associate deflation with depression. In their opinion, falling prices make it harder for debtors to service their debts. Widespread defaults on those debts could cause a financial crisis and a severe economic downturn. I believe their thinking is that simplistic.

Yet, perversely, the central banks have been responding to the risk of deflation with ultra-easy monetary policies in an effort to stimulate more debt-financed spending. It seems to be working in the US, though a significant portion of borrowing in recent years has been in the corporate bond market to finance stock buybacks. More recently, debt-financed M&A activity has picked up as well. Often, these deals are associated with cuts in payrolls.

On the other hand, debt is financing a US boom in multifamily housing construction and solid activity in commercial real estate. Mortgage-financed new and existing home sales also have continued to recover from the Great Recession. Auto loans are financing cyclical highs in auto sales.

US consumer price inflation measures excluding food and energy are closer to 2% than to zero. So the Fed is more relaxed about deflation and started raising interest rates again at the end of last year. However, Fed officials constantly and consistently say that the process of normalizing monetary policy will be very gradual.

In the Eurozone and Japan, both headline and core consumer inflation rates are close to zero, i.e., bordering on deflation. Ultra-easy money has failed to stimulate borrowing and demand in their economies. Both have been in growth recessions over the past few years. Since Q4-2010, real GDP is up 10.5% in the US but only 3.2% in the Eurozone and 2.7% in Japan.

The Chinese authorities don’t have a deflation problem as measured by their CPI. They do have one as measured by their PPI, which has been falling on a y/y basis for the past 51 months through May. They have been all too successful at pumping lots of credit through their banking system in the economy. Perversely, rather than fueling demand, much of it has financed more deflationary capacity expansion. China actually stands out as the one country where deflationary pressures may very well lead to lots of debt defaults and a severe downturn. However, easy money is exacerbating the problem rather than ameliorating it. Let’s dive into the widely feared sea of deflation and see what we can see:

(1) Durables deflation. There are seven countries that report the durable goods, nondurable goods, and services components of their CPIs. Here is what they’ve done since 1996:

Durables: US (-16.1%), Eurozone (-1.4), Japan (-50.1), Sweden (-31.1), Switzerland (-26.4), Taiwan (-28.9), and UK (-26.9).
Nondurables: US (51.1), Eurozone (37.1), Japan (9.1), Sweden (36.0), Switzerland (-13.7), Taiwan (51.0), and UK (37.7).
Services: US (73.7), Eurozone (48.3), Japan (5.8), Sweden (45.9), Switzerland (7.5), Taiwan (21.1), UK (96.3).

The pattern is obvious: There has been across-the-board deflation in durable goods, while nondurable goods and services have continued to inflate almost everywhere. The only difference recently (i.e., since mid-2014) is that energy prices have contributed deflationary pressures to the nondurable goods components. Durable goods deflation has been mostly good deflation, attributable to big gains in manufacturing productivity thanks to technological innovations. Weak demand hasn’t been the source of the deflation. On the contrary, durable goods deflation certainly has benefitted consumers around the world, lifting their purchasing power and standards of living.

(2) Inflating & deflating in the US. In the US, durable goods deflation has been widespread, though five of the remaining six countries have had more of it since 1996. The US has had the highest inflation in nondurable goods of the seven. It’s had the second-highest rate of services inflation.

Interestingly, the CPI inflation rate in the US has tended to exceed the PCED measure mostly because of their respective services components. In April, this component was up 2.7% in the CPI and 2.2% in the PCED. There has been a persistent gap between medical services in the CPI, currently up 3.1%, and in the PCED, up 1.3%. That’s because the former covers all out-of-pocket expenses, but does not include costs paid by government health care programs. The PCED covers both.

Furthermore, while rent inflation is measured identically in both the CPI and PCED, it has a much higher weight in the former (at 32%) than in the latter (at 15%). Rent has been among the categories with the highest (and rising) inflation rates during the current economic expansion.

(3) Yen devaluation fails to reflate. Among the most startling economic events since late 2012 has been the 38% depreciation of the yen through late 2015. The Japanese government implemented a stimulative program in 2013, widely known as “Abenomics,” that included massive QE by the BOJ. The aim clearly was to devalue the yen with the goal of ending deflation by boosting export volumes and import prices. It hasn’t worked. Exports did rebound for a short period, but during April were back down to the lowest pace since January 2014.

Inflation remains around zero. As shown above, Japan has had more durable goods deflation than other countries, and the least inflation in nondurables and services. In Japan’s case, weak demand certainly has contributed to overall deflationary pressures. Much of that can be attributed to the rapidly aging demographics of the country. There’s nothing the BOJ can do about that.

(4) Leak in ECB’s monetary pump. The ECB remains committed to providing more ultra-easy monetary policy to revive economic growth and inflation. Indeed, starting this month, the central bank will expand its QE program to include corporate bonds. Loans to the Eurozone’s private sector rose just 0.5% y/y through April, notwithstanding a 33% increase in the ECB’s balance sheet over this period and the introduction of negative interest rates on June 5, 2014. The problem may be that many borrowers borrowed too much during the previous decade’s boom, and don’t have the capacity to borrow more.

(5) Refinancing paradox. So why haven’t the historically low interest rates provided by the central banks revived growth and inflation? Too much accumulated debt stimulated by easy credit conditions in the past is weighing on demand. Aging demographics are also a drag on growth and inflation. Technological innovation is disruptive and inherently deflationary. Some deflation, especially in durable goods, should be welcomed as good inflation.

Most perverse is the likelihood that ultra-easy monetary policy has contributed to secular stagnation. It has allowed borrowers to refinance at lower rates. However, on balance, it seems that supply-side borrowers are benefitting more than demand-side borrowers. In other words, companies that might have been forced out of business or forced at least to reduce capacity continue to produce, hoping to sell to consumers who remain wary of borrowing more even at lower rates.

Of course, record-low interest rates may be forcing consumers to save more given the pathetic return on their assets. Furthermore, they have less income from their fixed-income portfolios to spend. This all adds up to persistent secular stagnation, with central banks continuing to provide easy monetary conditions in the hopes that their demand-side models will come back from the dead.

US Employment: Shortage of Jobs or Workers?

Payrolls rose just 38,000 during May, the weakest gain since September 2010, and the previous two months were revised downward by 59,000 collectively. However, something just doesn’t add up with these weak numbers. Consider the following:

(1) Record job openings. In March, total job openings rose to 5.76 million, nearly matching July 2015’s record high.

(2) Solid ADP payrolls. Private-sector payrolls as measured by ADP rose 173,000 during May, well above the measly gain of 25,000 reported by the Bureau of Labor Statistics. So far this year, the former is up 940,000 while the latter is up 699,000.

(3) Confidence survey upbeat. In the Conference Board’s monthly survey of consumer confidence, the percentage of respondents who said jobs are plentiful remained around recent cyclical highs at 24.3%.

(4) Earned Income Proxy still rising. Our Earned Income Proxy for total wages and salaries in the private sector rose 0.3% m/m during May to a new record high. It reflects the tiny increase in private payrolls as well as the 0.2% increase in average hourly earnings; the average workweek was unchanged. The EIP is up solidly by 4.2% y/y. (See our YRI-EIP.)

(5) Troubling signs. I think that in the worst-case scenario, the latest payrolls suggest that there may be a big skills mismatch between all those unfilled jobs and the available workers to fill them. More likely is that May’s weak report is misleading, though we are disturbed by the downward revisions in the prior two months. The 12-month sum of those revisions remains positive, but barely so--which in the past has been a harbinger of trouble for the labor market.

Furthermore, in May the percentage of industries reporting higher payrolls fell to 51.3% on a one-month span and 53.2% on a three-month span. Both are the lowest since 2010. May’s non-manufacturing PMI fell from 55.7 during April to 52.9 last month partly because its employment component fell from 53.0 to 49.7, back down at February’s reading, which was the lowest reading since February 2014.