About The Authors

Friday, April 6, 2018

Licensing our technology to China? Shhh ... it's nicely profitable

Mike Smitka, Economics
Washington and Lee University
Bass II, RCS

As an observer and sometimes participant in industrial policy debates for 30+ years, few industries are "winner takes all." It's just not the way either technology or competition works. In addition, few technologies turn out to be national in significance. Lots of players, lots of niches, not many jobs. There's typically a long delay between invention and innovation, and initial commercialization seldom makes money. Most patents are never used.

Furthermore, producers can't capture all the benefits. Users have their place in the value chain, except for the odd discrete product they're the ones who integrate it into what final consumers purchase. If the technology lets final purchasers reorganize their businesses, they too capture benefits. As a result, the whole world benefits from advanced countries subsidizing new technology. The gain to any one country is much smaller. And it's a big world: no one country can dominate more than a few slices of the market, even when they are the size of the US or China.

So here's a devils advocate position: the more China subsidizes their push into new (for them) technologies, the better it is for us, the United States. Let China bear the costs, which are upfront, and let us reap the eventual benefits. For the MBA's out there, think net present value: it's negative.

...Don't kill the goose!...

Oh, and licensing is a fool's game, and it's not us who are being fooled. Licensing is only good for yesterday's technology, you can't license something that doesn't yet exist or has no obvious value. In the auto industry, the product cycle means that a MY2018 vehicle represents the technology of 2014-2015 – and the technology in a Honda CR-V is going to lag that in the Acura RDX. Remember, too, that licensing is only the first step. The Chinese "partner" has to then get that technology into production. That takes a good 3 years. Licensed technology is old technology.

And licensing only covers what the other party can see – that's the equivalent of Tesla's purchase of lots of robots. Yes, you've the hardware and the access to the patents. But that doesn't tell you how to make a factory full of them operate, or how to design products for which automation saves money. That's tacit knowledge, embodied in part in teams of engineers who are in turn embedded in a larger engineering operation and networked with counterparts at component suppliers and capital equipment vendors. If the Chinese want to get their factories to run and their products to work right, they have to hire teams of consulting engineers. So those licensing get fees up front, get ongoing training contracts, and are the gatekeepers for buying specialized equipment and materials and so on. All of these can be nicely profitable, while leaving the Chinese perpetually 5 years behind.

This will vary from industry to industry. But many of the companies being "ripped off" are in fact crying all the way to the bank. They just have to be careful not to boast about it lest they kill the goose that is laying the golden eggs. How long has Shanghai GM been up and running? – the Buick Sail project started around 1999, the vehicle launched in 2001. So after 20 years how have they fared? They're still in the drivers seat for their regular trips to the bank.

Yes, GM's local partner, SAIC, is in the back seat. Why? Parts is because technology isn't readily transferred. But remember where the car is headed: SAIC is also on the way to the bank. Being a passenger isn't so bad, since GM delivers profits like clockwork. Neither SAIC nor the numerous other "domestic" Chinese auto companies have managed to make regular money on their own. Entering the auto industry is hard. Better to license, and license, and license some more. Don't kill the goose!

Tuesday, January 30, 2018

The Trump Bump in Coal

Mike Smitka
Washington and Lee University

I've no idea what Pres. Trump will say in his SOTU address, but it is useful for me, living near coal country, to look at what's happened in my neck of the woods. (BTW, I do live in woods!) In particular, can any president return jobs to regions with long-standing employment challenges? We in the US have tools to address "cyclical" unemployment, and hence affect what goes on at the national level. "Structural" unemployment is a different matter. Regional problems are, well, regional, with different origins. If we really want to do something, we need thus to diagnose causes and tailor solutions, region by region. No single national policy will do the trick. But in a Federal system we're not set up to do this: regional policy butts up against our many states, and a region may be comprised of slices of geography lying in several different states. Overcoming this institutional barrier – creating cross-state cooperation – isn't something possible without a great deal of hands-on, and likely heavy-handed, leadership.

Coal is a case in point. There's "met coal" that gets converted into coke and is used in smelting steel from iron ore. Imports aside, we (and everyone else) have become really good at recycling steel. The biggest example there is the rise of minimills, which recycle scrap into useable product. But even "basic steel" uses a charge of scrap that goes into the basic oxygen caldron alongside iron from the smelter. There are imports as well. The bottom line is that steel is a declining source of demand.

Then there's "thermal" coal that goes into power plants. Here the challenge is that alternatives are attractive. When oil prices are high, coal production does OK – but it doesn't rise, despite our growing economy. Today however energy prices aren't high, and coal output reflects that. Tax benefits aren't going to change that. (Note: The more important alternative today is natural gas. I tried to put natural gas prices into the chart, but it made the graph unreadable. Nor did I try to address how much coal is from WV versus Colorado or other regions where mining is less costly.)

Is there anything that a US president could do about this? Probably not: our current president is not to blame. But whatever might be done to help supporters in West Virginia won't help those in Ohio. There's no generic solution. Retraining programs might help, but even if community colleges could develop programs offering "portable" skills, those who finish programs would need to leave the region to find a job, so it might be helpful to add a one-time relocation allowance to training programs. That won't save the affected communities. It would help build a better future for the children of miners, and for younger Americans in general.

Saturday, December 23, 2017

America First: Federal Debt

Mike Smitka / Dec 23, 2017
Edited from a post on his Econ 102 course web site

Food for thought. The new tax bill will add $1 trillion in additional debt over the next decade – I emphasize that because the US already is adding debt faster than the economy is growing. The same is true of Japan. (Europe varies, but the OECD provides data that should allow an estimate of the growth rate of aggregate Euro Zone debt. I've not looked at their web site.)

As I will discuss in class during the Winter 2018 term, debt can't grow faster than the economy forever. The analysis is straightforward, though I won't do the arithmetic here. The stock of debt affects dynamics: the higher the level of debt, the more that higher interest rates make matters worse. So we've been helped since 2010 by exceptionally low interest rates. That is changing, with the Fed likely to bump short-term rates from 1.25% to 2.25% or higher in 2018. I expect rates to continue moving higher in 2019, once the new tax cuts phase in. For better or (in this case) worse, the Republicans in Congress are not conservatives, they're radicals who give only lip service to fiscal responsibility. At some point taxes must increase – again, debt can't grow faster than the economy indefinitely, and the deficit dwarfs the actual budget so even draconian cuts won't do the trick. (See the bottom graph, which uses Congressional Budget Office "cyclically adjusted" deficit estimates.)

Who on the political horizon is equal to the unpopular but necessary task of raising taxes? Japan's case does not make me optimistic. The issue is worse there, because debt is much higher, and thanks to demographics the economy is growing more slowly. Fiscal conservatives at the Ministry of Finance and elsewhere began talking about this issue in earnest in the early 1980s. Over time the need to cut deficits became sufficiently accepted across the political spectrum to be turned into legislation. The resulting shohizei (消費税, national consumption tax) took effect in 1989 with a modest 3% rate, as part of a broader tax reform that lowered other rates. The first big boost, to 5%, took effect April 1, 1997. The magnitude and timing of the increase was ill-judged, and the hike's initial impact was amplified by "buy now" campaigns for big-ticket items. Car sales boomed through late March 1997. Sales then fell off the cliff – I was living in Japan and visited a car dealership towards the end of the March. It was empty, because the registration process meant any sale wouldn't be recorded until after April 1. The consequent recession was such that, while the ruling Liberal Democratic Party continued to talk of further rate increases, it took 17 years to implement the next hike. There was always a good reason – the 9/11 US terrorist attack of 2001, the Lehman Shock of 2008, the 3/11 Tohoku earthquake in 2011, and pending elections all seemed to conspire to overcome good intentions. Meanwhile, and as expected due the to the rise of expenditures to cope with the aging of the population and various short-term stimulus and reconstruction projects, deficits remained stubbornly high. And that's in a country where the political consensus has remained one that taxes are too low, not that they are too high.

Tuesday, November 7, 2017

Cars are a long-term payment scheme

“The bottom line is that most people do not buy a home but rather commit to a long-term payment scheme.”

Izabella Kaminska, FTAlphaville, November 7, 2017

And the good phrasing doesn't stop there: her line refers to a Convexity Maven (Harley Bassman) note "House of Cards."

The above lede on the Financial Time's Alphaville blog caught my eye. It applies equally to the auto industry. Over 70% of new car purchasers use some sort of finance, and trade those wheels in before the loan matures for another financed "ride". At a 7.6% interest rate (the base rate from a major OEM's web site), and with a 7-year loan, a "purchaser" will still owe half the face value of the loan at the 4 year point. With 17% depreciation per year, their trade-in will just equal their loan value in the 52nd month.

Even I took out a loan – a 0.9% financing deal was too good to refuse. However, I didn't do a long loan, only 4 years, and I put 50% down so that I'd never be upside down. As it turns out, though, interest rates have stayed really low, and my short-term savings hasn't been earning at best 0.1%. So I might just as well have dipped into savings and paid cash for the whole thing. In any case, the four year consolation prize of being free of car payments will soon be here.

The point at which someone trapped in a long-term car payment scheme breaks even will vary with the depreciation rate, interest rate and length of the loan. Here are sample calculations. Note that a shorter-maturity loan pays down much more principle in the first year, and so purchasers aren't underwater for long. The cost, of course, is a higher monthly payment. Here I assume (unrealistically) that the loan is for 100% of the car value as reflected in the resale market – no extended service plan rolled in, no downpayment. The number of permutations with the latter is unwieldy. So if the car is valued at $60,000, then with a $10,000 downpayment (or tradein plus incentives), an 84-month, 7.6% loan remains rightside-up throughout but provides $5000 equity only in the first year, before that $60K has depreciated much, and then again from the 55th month.

Car payment calculations, $50,000 loan
Upside down

Monday, October 30, 2017

Graph(s) of the Day: Govt Transfer Programs

Mike Smitka

Ten days since the last graph ... time to add another. Here are the largest transfer programs in the US as a "stacked graph." These include those not only of the Federal government but also much smaller state and local programs. I've not included private transfers. Those aren't large, or more precisely are no longer large, in an era when few companies provide employer-funded defined-benefit pensions.

Here are these data normalized to GDP - the vertical axis should be "per cent" but I can't relabel so you have to note .14 is 14%. As expected, the increase looks much less dramatic, and you can see the sharp falloff of unemployment insurance as the labor market improved post-2011.

First, you can see the magnitude of government transfer programs. In our $20 trillion economy, they come to $2,700 billion or about 14% of GDP. (Remember, transfers don't show up directly in GDP, they are personal income and enter the economy only when someone spends the money.) However, the graph is not particularly helpful, because this is nominal GDP, which reflects not just "real" growth but also inflation. To better convey the magnitude, you should divide each of these series by nominal GDP. Rather than ask you to do that – it is straightforward task in FRED – here's the graph for the total divided by nominal GDP. You can see that there was a big increase with the Great Recession, but that the amount is slowly decreasing. I expect that is muted because of the retirement of the Baby Boomers, and not just by the still partial recovery during which 2% of prime-aged workers have yet to re-enter the labor force. But admit that I have not dug in deeper.

We can also look at the composition. In the early 1950s there were few retirees as a share of the population, and many of those retired did not qualify for social security. That changed as more and more workers were "vested" in the program (which was enacted only in 1937). As a share unemployment was much higher, but that's as part of a very small total. The catch-all other category looms large as a share of transfers, but small as a share of GDP. Neither Medicare nor Medicaid existed until 1966. Today retirement programs predominate: social security and medicare together are $1.6 trillion. State-level healthcare programs (Medicare) adds another $560 billion. The remainder includes miscellaneous small items such as veteran's benefits ($90 billion), food stamps (or formally SNAP, at $65 billion), and tax rebates, such as the $7,500 that a well-heeled American gets back from the purchase of a Tesla Model S but also the much more substantial earned income credits that help a low-income individual offset various taxes. In other words, what we would commonly refer to as "welfare" is too small relative to our $20 trillion economy for FRED to break them out as separate line items. [For much, much more detail see Table 3.12 Government Social Benefits in the National Income and Product Accounts.]

All of this provides an anchor to the economy: these flows continue even if we enter into a recession. That's important, because when someone loses a job, the impact spills over to local retailers and others, and leads to additional job losses, the "multiplier" of Economics 102. Now as you can see from the data, unemployment benefits are generally trivial, and they are also time-limited (24 weeks or roughly 6 months). But they increase sharply during a recession, and act as another component of our "automatic stabilizer," allowing people to cut consumption less and thereby partially offseting the spillover effects from the loss of a job. It's important not to overstate the cushion this provides, as it's hard to find a job during a recession. Benefits aren't very generous, either, averaging only about $300 per week – for many workers the replacement rate is only about 50%. Given the severity of the Great Recession, Congress did recognize the importance of even partially stemming the cascading effects of job losses; the ARRA temporarily increased the duration of unemployment insurance for up to 99 weeks. If you'd been earning $36,000 per year, however, and just barely making car and house payments, you couldn't make ends meet on $1,200 per month in unemployment checks. Goodbye house, but keep the car or you'd be unable to work even if a job became available.

Wednesday, October 25, 2017

Sticky Prices and Macroeconomics

Mike Smitka

We're now starting on AD/AS analysis in my intro macro class. It involves a lot of handwaving, given the simplifications imposed by using blackboard models – do we want our y-axis to be "the" price level or the rate of inflation π or (later in the term) the rate at which π is changing? At the same time the x-axis is a broad aggregate, (real) GDP, that we conceptualize as a physical aggregate. So why do they behave as supply and demand curves ought to, sloping up and down, respectively? One answer is that prices are sticky. If you want the argument, go to a textbook.

The more general issue is one that plagued the classical economists: if prices adjust freely, then why should there ever be unemployment, or other unsold goods? Why should productive capacity in an economy ever sit idle? Here the key is that prices are sticky. My wage, for example, is set but once a year, and that in accord with a salary pool based on expected tuition and other revenue sources, which likewise are updated only once a year. So if a shock hits the economy – prices rise unexpectedly (or inflation, the rate of price rises, rises unexpectedly) I will see my pay fall behind, and even if the university does set out to adjust things, it may take a couple years to get back in synch. Since labor accounts for a bit over half of all costs for producers of goods and services in the US, most attention focuses on that behavior.

Now I've never seen a breakdown of the frequency of wage changes across the economy, though I've never looked and I'm not a labor economist. But the Federal Reserve Bank of Atlanta Inflation Project does recompile the data that underly the Consumer Price Index into goods and services that are "sticky" and those that are "flexible". Over time they average out to be about the same, inflation truly is general, but in the time frame of a few years they can and often do move in opposite directions.

Their web site details which goods are sticky and which are flexible, with the weights each have in the CPI basket of goods and services. On the flexible end are things such as gasoline, electricity prices, food purchased in a supermarket, women's clothing and (yes!) new car prices – think of all those rebates and subvented [jargon for manufacturer-subidized] loans and leases that appear when sales are slow. On the sticky side are personal care services (how many years since my barber upped his price? – the sign's yellow!), fast food prices, school fees and beer prices.

For building a supply model we'd want to look at components of business costs besides labor, things like the billing rates for lawyers and accountants or xerox paper. No one to my knowledge has put together a comparable flexible/sticky producer price index. But for building theory, all we need to know is that in the real world prices are sticky.

Saturday, October 21, 2017

Dealer Musings (II): Ruggles on Gap Insurance

David Ruggles

Gap insurance is a by-product of the extension of car loans to longer maturities. Once only 3-4 years, loans now extend out to 7 years (and sometimes more). At current interest rates of 7.6% (quoted on a Chevy dealer website), at the end of 3 years the purchaser will have only paid $18,200 on a $50,000 loan. However, the resale value may only be $25,000 or 50%. So the purchaser will be “upside-down” on their loan by $7,000. Now there’s generally some sort of downpayment, but purchasers can add on extras that aren’t figured into the insurance value of the car but might be built into the loan. So if they total their car and are paid only the replacement cost (less deductible!), they’re in trouble. Or in this case, their truck in Texas was underwater, and their auto loan is, too.

Now consumers can be quite naive about this risk. That's not true of banks. If the borrower defaults and the bank repossesses the car, they’ll be out the “gap” between the balance of their loan and the wholesale value of the car (plus the out-of-pocket costs of hiring a “repo” specialist, and cleaning up and transporting the vehicle to auction). Hence “gap” insurance. (Of course the same issue faces anyone involved in vehicle finance.)

...don’t be surprised if some GAP insurers become insolvent...

Friday, October 20, 2017

Friday class, Friday graph of the day: Trade

Mike Smitka

Krugman (co-author of Krugman & Wells, our Economics 102 textbook) won his Nobel Prize for work in international trade, but his "open economy" chapter remains at the end of the book. That textbook structure dates back to the first true "principles" text in 1947, written by Paul Samuelson (another Nobel laureate, and a co-founder with Heckscher and Ohlin of modern trade theory). Since then all successful new texts chose to market themselves as potential replacements, which is realistic only if they are close to what econ profs were already using. We do not see product differentiation! But that's a topic for Econ 243 (Economics of Strategy / Industrial Organization) and not Econ 102.

Thursday, October 19, 2017

Dealer Musings (I): Ruggles on Recalls

David Ruggles

I wish I had something pleasant to say about things affecting dealerships. I'll try to come up with something before the end of this series. But if a person long on the sales side of the industry can't think up an optimistic story, maybe there isn't one. After all, how many perennially downbeat sales people have you met?I start though with safety recalls, and follow with gap insurance and then the latest twist in branding that pits the Factory against dealers.

Wednesday, October 18, 2017

Graph of the Day: Wed 18 Oct

Mike Smitka

Today saw the latest release on residential construction. The default graph on FRED is the total number, of course with multiple series for new starts versus permits versus under construction, and by region. But between 1960 and today the US population rose 80%, from 180 million to an estimated 325 million. Everything else being equal – and it's not, we know that the baby boomers are retiring and downsizing – we would expect the desired amount of housing to reflect the size of our population. So my graph makes that adjustment, dividing by population. As you can see, the current level of housing under construction has doubled since its December 2012 trough, but remains significantly lower than at any time prior to the Great Recession.