The Importance of Worker Productivity Improvement

Worker productivity is an important factor in understanding the economy in the United States (and for other economies). Worker productivity, as defined by the U.S. Bureau of Labor Statistics, is the output per hour for workers in nonfarm businesses. It has a major impact on the need for workers as the economy grows. For example, if the economy grows at 2% and worker productivity does not grow, then theoretically 2% more workers are needed. On the other hand, if worker productivity grows at 2%, then there would be no need for additional workers. Here the assumption is that the need for workers is directly proportional to the growth in the economy.

This chart shows average yearly labor productivity and the yearly average percent change in worker productivity over the past 30 years (from the Bureau of Labor Statistics, https://www.bls.gov/#productivity). Also, included in the graph is the labor unemployment rate over the same period. The data shows that overall worker productivity has grown throughout the period. The rate of change increased dramatically in the late 1990s, followed by a decline in the rate of change in the 2002-2010 period. Since 2012, the rate of change in worker productivity has been consistently less than 1.5%. The blips in the rate of change during 1990-1992, 2000-2001, and 2009-2010 were due to economic recessions that occurred during these periods.

 

To get the full picture, unemployment rate is shown on the graph. Note how unemployment has significantly dropped since 2010. As discussed, other than higher productivity, another way to keep up with the need for more output is to use more workers. The analysis shows how as the rate of productivity increases, unemployment is reduced. The implication is that the output shortfall is being made up by hiring more workers. Of course, this can only keep happening for a limited time. The concern is that the limited labor force may begin to impact output in the United States. This could be a major economic problem for the United States resulting in a capacity shortfall in the future. This can also lead to an increase in inflation due to higher wages as companies compete for the smaller pool of unemployed workers.

Consider the following example. Assume that next year the Output (think about his as the demand) in our industry is expected to go up 5% from 100 units to 105 units and Labor Productivity is expected to increase by 2%. If the workforce stays at the same level, they could only produce 102 units due to the productivity improvement. In order to reach the full potential of the industry, the extra 3 units need to be produced in some way. There are several ways that this could be done. One might be to work overtime, another might be to hire more workers from the pool of workers that are currently unemployed. If the current unemployment rate is 5%, then if 95% of the workers can produce 102 (=100 x 1.02) units, approximately 97.7% (=.95(105)/102) of the workers would be needed to produce 105 units. This would result in an unemployment rate of less than 3%, which is probably not realistic. It is likely that the workforce might limit potential growth. Also, all the pressure to hire those extra workers might result in some inflation due to the higher wages needed to attract those workers.

To take this a step further, one needs to consider trends in the size of the workforce. This needs to look at trends in the number of people that are looking for work and basic changes in the number of people entering and leaving the potential workforce pool. The Bureau of Labor Statistics tracks data on the Labor Force Participation Rate and the Employment-to-Population Ratio and this data is shown in the following chart.

The Labor Force Participation Rate has held steady at a little less than 63% since 2014. The Employment to Population Ratio has increased from 58% to 60% since 2010. This is promising, but the key to whether output will be choked is the actual number of workers that will be available to work. This, of course, depends on the shifting population demographic in the United States.

A projection made by The Bureau of Labor Statistics is shown in the following table. Their data forecasts that the Labor work force will increase by 6.6% from 159,187,000 to 169,650,000 workers between 2016 and 2026. This is an annual growth rate of only .6%. This projection considers the change in mix of workers and the growth in the U.S. population. Note that the percent distribution of workers 55 and older moves only from 22.4 to 24.8% between 2016 and 2026, so much of the “aging” of the United States population has already occurred, at least from the standpoint of workers. Disturbing is the projection that the annual growth rate in total workers will only be .6%.

Civilian Labor Force 2006, 2016, 2026 (U.S. Bureau of Labor Statistics)

Group	Level (1,000)			Change		Percent Change		Percent Distribution			Annual Growth (%)
	2006 (actual)	2016 (actual)	2026 (forecast)	2006-16	2016-26	2006-16	2016-26	2006	2016	2026	2006-16	2016-26
Total, 16 years and older	151,428	159,187	169,650	7,759	10,463	5.1	6.6	100.0	100.0	100.0	0.5	0.6
Age, years:
16 to 24	22,394	21,202	19,868	-1,192	-1,334	-5.3	-6.3	14.8	13.3	11.7	-0.5	-0.6
25 to 54	103,566	102,248	107,634	-1,318	5,386	-1.3	5.3	68.4	64.2	63.4	-0.1	0.5
55 and older	25,468	35,737	42,148	10,269	6,411	40.3	17.9	16.8	22.4	24.8	3.4	1.7

 

Going back to the previous example, assume that next year the demand in our industry goes up 5% from 100 units to 105 units, Labor Productivity increases by 2%, and the workforce goes up by .6%. In this case, the workforce can produce 102.612 (=100 x 1.006 x 1.02) units due to the increases in workforce size and productivity gains. The remaining 2.388 units need to be produced by overtime or other technology innovations. Given this scenario approximately 97.2% (=.95(105)/102.612) of the workforce will be needed. The unemployment rate would be less than 3%.

It sounds to me like a lot of people are going to be working overtime in the future!

 

 

F. Robert Jacobs 12/6/2018

 

Innovative E-commerce Software from ShippingEasy

by F. Robert Jacobs, Ph.D.
Oct 18, 2018
ShippingEasy is an innovative company that offers software for managing many common supply chain tasks. The software features logic that ties users into the best commercial shipping rates offered by the U.S. Postal Service, United Parcel Service, and FedEx. As orders arrive in shopping carts, platforms, and marketplaces the software automatically plans inventory picking and shipments to efficiently process the order.

ShippingEasy has developed a “skill” for Alexa the cloud-based voice service run by Amazon. Using the voice-driven capability, orders and shipments can be managed with simple voice commands. The simple conversational user interface speeds the training of workers and improves overall productivity.

The Alexa/ShippingEasy collaboration is a great example of how technologies from different companies can be integrated to create more capable systems. In this case the Alexa speech recognition technology is programmed to automatically generate commands for the ShippingEasy e-commerce software, thus replacing the tedious computer input user interface. These types of integrated technologies will be a key feature of the intelligent systems of the future.

Tesla Appears to be Preparing to Build a Plant in China

According to a WSJ article published on Oct 17, 2018 (Tesla Advances in China, Buying Land for a Factory), Tesla has agreed to by a 210-acre site near Shanghai for $140 million.

• Tesla said it was “accelerating the construction of our Shanghai factory” in response to the U.S.-China trade dispute. Prior to July 2018 China levied at 15% tariff on all imported vehicle.
• In July 2018 China introduced an additional 25% tariff in retaliation for new U.S. auto tariffs. This raises the tariff on autos imported into China to 40%.
• Tesla sold about 17,000 cars in China in 2017, roughly 50,000 in the U.S. and 103,000 globally.
• Tesla’s cars cost 55-60% more in China that electric cars built in China, according to China.
• The company intends to eventually produce up to 500,000 cars a year in a plant on the site.

Tesla Financial Predictions Based on Learning Curve Analysis

by F. Robert Jacobs, PhD.   10/16/2018

Given the estimates from the learning curve analysis of the production capability of Tesla, it is interesting to predict what the financial implications might be. The following table has the financial data reported by Tesla in quarterly reports since 2016-Q4.

Quarterly Financial Data (all revenue and expenses in thousands

	2018-Q2	2018-Q1	2017-Q4	2017-Q3	2017-Q2	2017-Q1	2016-Q4
Automotive revenue	$3,357,681	$2,735,317	$2,702,195	$2,362,889	$2,286,616	$2,289,600	$1,994,123
Automotive gross margin – GAAP	20.6%	19.7%	18.9%	18.3%	27.9%	27.4%	22.6%
Expenses	$2,665,999	$2,196,460	$2,191,480	$1,930,480	$1,648,650	$1,662,250	$1,543,451
Model S and X units delivered	22,300	21,800	28,320	25,116	22,000	25,000	22,299
Model 3 units delivered	18,440	8,180	1,550	220
Total units delivered	40,740	29,980	29,870	25,336	22,000	25,000	22,299

Using the data, we first developed a regression model to estimate the revenue from Model S and X products (these are grouped together, since they are similar), and the Model 3 product. The regression indicates that the average revenue for Model S and X is $94,045.54, and that the average revenue for Model 3 is $70,610.36. The F statistic for this model is 1433 indicating very high confidence in the numbers.

Similarly, a regression analysis was run on costs over the period. Fixed expenses over all three models were estimated together with variable expenses for the combined S and X product, and for the Model 3 product. The model indicates that fixed expenses are approximately $18,429,936.85 per quarter. Expenses for the Model S and X product are $72,033,599.15 and for the Model 3 are $59,599,668.52.

Using this data, the following table was prepared that estimates Revenue, Expenses, and Net Profit for 2018-Q3 and beyond.

Quarter	Model S and X units delivered	Model 3 units delivered	Revenue	Expenses	Net Profit
2016-Q4	22,299	0	$1,994,123	$1,543,451	$450,672
2017-Q1	25,000	0	$2,289,600	$1,662,250	$627,350
2017-Q2	22,000	0	$2,286,616	$1,648,650	$637,966
2017-Q3	25,116	220	$2,362,889	$1,930,480	$432,409
2017-Q4	28,320	1,550	$2,702,195	$2,191,480	$510,715
2018-Q1	21,800	8,180	$2,735,317	$2,196,460	$538,857
2018-Q2	22,300	18,440	$3,357,681	$2,665,999	$691,682
2018-Q3	26,903	55,840	Not reported yet
Forecast
2018-Q3	Actual from above	$6,472,990	$5,284,395	$1,188,594
2018-Q4	24,217	78,948	$7,852,069	$6,468,159	$1,383,910
2019-Q1	24,217	128,874	$11,377,349	$9,443,721	$1,933,628
2019-Q2	24,217	187,010	$15,482,339	$12,908,596	$2,573,743
2019-Q3	24,217	259,051	$20,569,234	$17,202,261	$3,366,973
2019-Q4	24,217	346,102	$26,715,932	$22,390,469	$4,325,464

To explain this table, the numbers for 2016-Q4 thru 2018-Q2 are actual as reported by Tesla. For 2018-Q3, Tesla has reported delivery numbers, but the financials have not been reported. These estimates assume that Tesla will actually deliver the Model 3 cars that the Learning Curve estimates that they can make (apparently this has been a problem recently). In addition, these estimates assume that the Model S and X deliveries are equal to average past deliveries.

The implications are that potentially Tesla Automotive related Net Profit could more than triple by 2019-Q2, compared to 2018-Q2. Being realistic, the numbers beyond 2019-Q2 will probably not be realized since by that time the Model 3 order backlog should be gone, and regular market demand will dominate.

It will be interesting to see how this works out in the future.

Tesla Model 3 Learning Curve Analysis

by F. Robert Jacobs

This study is a Learning Curve[i] analysis of Tesla Model 3 production using data provided by the company[ii]. Using Model 3 production data from 2017-Q3 through 2018-Q3, the Wright learning curve function was fitted. The squared error (difference) between the predicted cumulative days to achieve cumulative production amounts and the reported cumulative production amounts was the criteria used in the fitting.

The two parameters fitted are the learning curve rate, and time for the first unit produced. The analysis indicates that the learning curve rate that Tesla has experienced is approximately 60 percent and that the time for the first unit was approximately 14 days.

The following graph shows the actual versus the predicted time/unit.

As can be seen the fit looks good. The F statistic comparing these two sets of values is only 1.18 which indicates they are not significantly different.

Using the model to predict future Model 3 production indicates that Tesla may be able to work off the order backlog by the end of 2019-Q2.

The most recent estimate of the Model 3 order backlog was 420,000 units. We see that Tesla cumulative production could well exceed this amount by the end of 2019-Q2. If one assumes that the Model 3 is being made on two assembly lines that run 16 hours a day, 5 days a week and 13 weeks/quarter, then the cycle time would be down to approximately 2 minutes per unit at the end of 2019-Q2. This would appear to be feasible. At this point, market demand would dictate production rates for the future and the learning curve rates would not be applicable.

---

[i] Wright, T. P., “Factors Affecting the Cost of Airplanes,” Journal of Aeronautical Sciences, 3(4)(1936): 122-128.

[ii] Tesla has provided this data in press releases on Oct 2, 2018, Jul 2, 2018, Apr 3, 2018, Jan 3, 2018, etc. The financial data are from Quarterly “update letters from Tesla. The data is available from http://ir.tesla.com/press-releases.

Tesla–A different mix of employees compared to other car companies.

Elon Musk, the savvy CEO of Tesla Motor, has recently posted on Twitter his need for “hard-core software engineers” to work at his company.  His electric-car company is looking to add thousands of employees in coming years.  The company is in a major development effort to write software needed for its future autonomous vehicles.

Tesla’s effort, known as Autopilot, is causing it to add employees faster than typical auto makers with such small volume.  Mr. Musk’s strategy is founded in developing this capability in-house rather than relying on outside firms.  In addition, to the major software development effort, Tesla makes cars and all the major components in plants located in southern California that were once owned by Toyota, General Motors and Solyndra.  Tesla even makes seats in an in-house production facility, something that is not common in the auto industry as other automobile manufacturers normally outsource this.

All of this in-house capability has resulted in Tesla’s ratio of employees to vehicle production to be be high compared to competitive luxury car makers.  Jaguar Land Rover recently was expected to sell about 500,000 vehicles and has 36,000 employees, a ratio of 13.8 vehicles per employees.  In comparison, Tesla is looking to sell 52,000 vehicles and employees about 14,000, a ratio of 3.7 vehicles per employee. 

Tesla is expecting much higher sales in the future.  The higher sales expectation combined with a strategy of keeping functions in-house makes Tesla appear much less efficient that competitors. 

The Top Dozen Supply Chain Innovations of All-Time

First Thoughts  

  By Dan Gilmore – Editor-in-Chief  

  April 23, 2015  

  The Top Dozen Supply Chain Innovations of All-Time

I am very keen on the subject of supply chain innovation right now, for a variety of reasons.

First, CEOs across the globe are putting innovation as the top or near top priority for the companies they run. It’s pretty simple why: innovation is what drives the growth and attractive margins. Supply chain in turn obviously has a key role to play in overall corporate/product innovation. This is perhaps most manifestly seen right now in the Internet of Things and how industrial companies especially are currently or prospectively leveraging IoT to create new products and services.

So, supply chains need to both support product innovation, while also innovating in its own domain. But I would argue that supply chain innovation is not well understood. Certainly we don’t have any good ways to measure it.

I have been doing some light collaboration on supply chain innovation with Dr. Jim Rice of MIT, who has also been researching this topic. MIT in fact has something it calls the Forum for Supply Chain Innovation, which is doing research in this area, among other activities.

Hope to have something more to share from the two of us soon on this, but there are a number of issues to be explored, from defining supply chain innovation, to measuring it, to the difference between innovation and continuous improvement and many more.

I am quite excited to be working on these types of questions, which I think will be good for the industry if we can wrestle some of them to the ground.

As some readers may remember, back in 2010 I came up with the top 10 supply chain innovations of all-time. Given my focus on this topic here in 2015, I thought it would be appropriate to repost that list, which I think is a good one. But too add some new value I decided to add two innovations to make it an even dozen.

I will note Rice used my list during his own excellent presentation on supply chain innovation at the MHI annual conference last October.

It was difficult to create this list, because many innovations either have no clear origins or came from a sort of a combined evolution along a number of fronts. This is especially true in terms of much supply chain technology innovation.

So then as now, I was looking for:

• Innovation for which we can identify pretty clearly that some single company or individual(s) was/were responsible for the breakthrough – though of course everything builds off the past

• The innovation had a deep and lasting impact on supply chain practice

That said, here is the expanded list, in reverse order:

12. The First True Network Optimization System: While there there were a few very primitive single commodity network optimization attempts by various academics that were not of much value, the first true “multi-product” network optimization tool was developed by Dr. Arthur Geoffrion and Dr. Glenn Graves, both of UCLA, in 1972. It was formally reported in the literature in a seminal work that appeared in the Management Science journal in 1974. That article is studied by many OR students even today. The network analyzed was that of Hunt-Wesson Foods. Geoffrion and Graves became two of the five co-founders of Insight, which still does this kind of work today, and really created the network optimization industry.

11. The Kiva Robotic Picking System: The idea for the orange AGV-like robots that bring inventory to order pickers was first conceived in 2003 by CEO Mick Mountz, and with the help of some MIT professors Kiva brought the technology to market less than two-years later. In 2012, Amazon spent an amazing $775 million to acquire Kiva – a small company at the time – which is one measure of the system’s value, as Amazon now keeps the technology to itself and eventually into dozens of its fulfillment centers. This was true innovation, and has ushered in the “goods to picker” concept that is now so prominent in materials handling circles.

No. 10: Taylorism: In the late 1800s, the great Frederick Taylor takes the first scientific approach to manufacturing. In the early 1880s, he invents the concepts of using time studies on the factory floor, and based on that work, the notion of “standard times” for getting specific tasks done. Later develops the concept of incentive systems and piece-rate pay plans. Taylor’s ideas were simply seminal – and often controversial – and dramatically influenced the practice of manufacturing over the next few decades and even to this very day.

No. 9: 3M’s Transportation Load Control Center: In 1982, 3M, like every other company, had to leave transportation decisions to each plant and distribution center. Roy Mayeske, at that time the Executive Director of 3M Transportation, had the idea to centralize transportation planning to look for network synergies. 3M took mainframe software that had been developed for Schneider National – one of its major carriers – and had it modified it to be workable from a shipper perspective. Ship sites called in planned shipments, and then carriers and routings were phoned back. The LCC is now of course a standard logistics practice today.

No. 8: Distribution Requirements Planning (DRP): In the late 1970s, Andre Martin ran operations for Abbott Labs Canada, and found himself caught between manufacturing and distribution managers, who could never seem to get inventory questions right and always blamed each other. Realizing that what was needed was a sort of Manufacturing Resources Planning for inventory distribution, Martin led a successful effort to build the first computerized DRP system, which in turn led to a book that created the software category of DRP, as several technology firms built products based on these ideas. Was in many way the start of today’s supply chain planning software industry.

No. 7: The FedEx Tracking System: After re-inventing the category of express parcel shipments, FedEx went a step further in the mid-1980s with its development of a new computerized tracking system that provided near real-time information about package delivery. Outfitting drivers with small handheld computers for scanning pick-ups and deliveries, a shipment’s status was available end-to-end. The FedEx system really drove the idea that “information was as important as the package itself,” and was foundation of our current supply chain visibility systems and concepts.

No. 6 – The Universal Product Code: Though the idea to use some form of printed and even wireless automatic product identification had been around for decades, lack of standards had precluded individual ideas from gaining any sort of critical mass. In 1970, a company called Logicon wrote a standard for something close to what became known as the Universal Product Code (UPC) to identify via a bar code a specific SKU, an effort that was finalized a few years later by George Laurer at IBM. The first implementation of the UPC was in 1974 at a Marsh’s supermarket in Troy, OH north of Dayton. The invention triggered the auto ID movement, forever changing supply chain practice and information flow.

No. 5: The Ford Assembly Line: Henry Ford actually got the idea for the assembly line approach from the flow systems of meat packing operations in the Midwest, but it was Ford’s adoption of the production approach with a continuously moving line for Model T’s in 1913 that revolutionized not only automobile assembly but took the practice of manufacturing to new levels in other sectors as well. Total time of assembly for a single car using the production line fell from 12.5 labor hours to 93 labor minutes, ultimately making cars affordable for the masses, changing not only supply chain but society.

No. 4: Economic Order Quantity (EOQ): Economic Order Quantity is a mathematical approach for determining the financially optimal amount of product to order from suppliers based on inventory holding costs and ordering costs. The original concept is generally credited to Ford Whitman Harris, a Westinghouse engineer, from an article in 1913, but it was a much later article in the Harvard Business Review in 1934 by RH Wilson that made EOQ mainstream. The formulas are still taught today, and the basis for much supply chain decision-making even in this era.

No. 3: The Ocean Shipping Container: It is hard to imagine today, but until the mid-1950s, there was no standard way to ship products on ocean carriers, and most were shipped on whatever container or platform the producing company deemed best. The result was terribly inefficient handling on both sides of the equation, poor space utilization on the cargo ships, freight damage, and high logistics costs. Enter Malcom McLean, legendary logistics entrepreneur and visionary who invented the standard steel shipping container first implemented in 1956 at the port of New Jersey. Someone would have thought of it eventually, but McLean’s invention started the explosion in global trade.

No. 2: P&G’s Continuous Replenishment: Until 1987 or so, order patterns in the consumer goods supply chain were almost totally dependent on whatever the manufacturer’s sales person and retail buyer decided between them. That’s until Procter & Gamble bought a mainframe application from IBM for “continuous replenishment” (which had been deployed a handful of times in other markets), re-wrote it for consumer goods to retail, and as a result dramatically changed that entire value chain by driving orders based on DC withdrawals and sales data.

P&G first implemented the approach with Schnuck’s Markets in St. Louis, with dramatic results in both lowering inventories while increasing in-stock at retail. KMart was next, taking pipeline diaper inventories from two months to two weeks – but KMart never completely embraced the possibilities. A legendary 1988 meeting between P&G’s CEO and Sam Walton led to a CR program there and changed supply chain history, helping propel Wal-Mart to retail dominance and providing the foundation for Efficient Consumer Response (ECR), Category Management, Continuous Planning, Forecasting and Replenishment (CPFR), and more.

  And finally…. (drum roll and envelop please):

No. 1: The Toyota Production System: When James Womack and several co-authors wrote “The Machine that Changed the World” in 1990, it was of course not a Toyota car that had such an impact, but rather the Toyota Production System (TPS) that was the foundation of the company’s dramatic success across the globe. Pioneered by Toyota’s Taiichi Ohno and a few colleagues, TPS not only is the foundation for today’s Lean manufacturing and supply chain practices, but the concepts have penetrated versus every area business. TPS truly did change the world.