Checkerboard Puzzle, Moore's Law, and Growth Prospects

To contact us Click HERE
My father the mathematician first posed the checkerboard puzzle to me back in grade-school, perhaps on some rainy Saturday. His version of the story went something like this:

The jester performs a great deed, and the king asks him how he would like to be rewarded. The jester is aware that the king is a highly volatile individual, and if the jester asks for too much, the king might just kill him then and there. The jester also knows that the king views his promise as sacred, so if the king says "yes" to the jester's proposal, then the king will honor that promise. So in a way, the jester's problem is how to ask for a lot, but have the king at least initially think it's not very much, so that the king will give his consent.

So the jester clowns around a bit and then says: "Here's all I want. Take this checkerboard. On the first square, but one piece of gold. On the second square, two pieces. On the third square, four pieces, and on the fourth square, 8 pieces. Double the amount on each square until you reach the end of the checkerboard."

In the story, the king laughs at this comic proposal and says,  "Your great deed was so wonderful, I would have happily done much more than this! I grant your request!"

But of course, when the king starts hauling up gold pieces from the treasury, he will discover that 2 raised to the 63rd power, the final spot on the checkerboard requires about 9 quintillion gold pieces (that is, 9 followed by 18 zeros). 

I've had some sense of the power of exponential growth ever since.  But what I hadn't thought about is the interaction of Moore's Law and economic growth. Moore's Law is of course named for Gordon Moore, one of the founders of Intel, who noticed this pattern back in 1965. Back in the 1970s, he wrote a paper that contained the following graph showing how much it cost to produce a computer chip with a certain number of components. Here's his figure. Notice that the numbers of component on the horizontal axis and the cost figures on the vertical axis are both graphed as logarithm (specifically, each step up the axis is a change by a factor of 10). The key takeaway was that the number of transistors ("components") on an integrated circuit was doubling about every two years, making computing power much cheaper and faster.


Ever since I started reading up on Moore's law in the early 1980s, there have been predictions in the trade press that it will soon reach technological limits and come to and end. But Moore's law marches on: indeed, the research and innovation targets at Intel and other chip-makers are defined in terms of making sure that Moore's law continues to hold for at least awhile longer. Stephen Shankland offers a nice accessible overview of the current situation in an October 15, 2012, essay on CNET: ""Moore's Law: The rule that really matters in tech"  (The Gordon Moore graph above is copied from Shankland's essay.)

As Shankland writes: "To keep up with Moore's Law, engineers must keep shrinking the size of transistors. Intel, the leader in the race, currently uses a manufacturing process with 22-nanometer features. That's 22 billionths of a meter, or roughly a 4,000th the width of a human hair." He cites a variety of industry and research experts to the effect that Moore's law has at least another decade to run--and remember, a decade of doubling every two years means five more doublings!

It's hard to wrap one's mind around what it means to say that the power of microchipo technology will increase by a factor of 32 (doubling five times) in the next 10 years. A characteristically intriguing survey essay  from the January 10 issue of the Economist on the future of innovation uses the checkerboard analogy to think about the potential effects of Moore's law. Here's a comment from the Economist essay:

Ray Kurzweil, a pioneer of computer science and a devotee of exponential technological extrapolation, likes to talk of “the second half of the chess board”. There is an old fable in which a gullible king is tricked into paying an obligation in grains of rice, one on the first square of a chessboard, two on the second, four on the third, the payment doubling with every square. Along the first row, the obligation is minuscule. With half the chessboard covered, the king is out only about 100 tonnes of rice. But a square before reaching the end of the seventh row he has laid out 500m tonnes in total—the whole world’s annual rice production. He will have to put more or less the same amount again on the next square. And there will still be a row to go.

Erik Brynjolfsson and Andrew McAfee of MIT make use of this image in their e-book “Race Against the Machine”. By the measure known as Moore’s law, the ability to get calculations out of a piece of silicon doubles every 18 months. That growth rate will not last for ever; but other aspects of computation, such as the capacity of algorithms to handle data, are also growing exponentially. When such a capacity is low, that doubling does not matter. As soon as it matters at all, though, it can quickly start to matter a lot. On the second half of the chessboard not only has the cumulative effect of innovations become large, but each new iteration of innovation delivers a technological jolt as powerful as all previous rounds combined."

Now, it's of course true that doubling the capacity of computer chips doesn't translate in a direct way into a higher standard of living: there are many steps from one to the other. But my point here is to note that many of us (myself included) have been thinking about the changes in electronics technology a little too much like the king in the checkerboard story: that is, we think of something doubling a few times, even 10 or 20 times, and we know it's a big change, but it somehow seems within our range of comprehension.

But when something has already been doubling every 18 months or two years for a half-century--it is continuing to double!--the absolute size of each additional doubling is starting to get very large. I lack the imagination to conceive of what will be done with all this cheap computing power in terms of health care, education, industrial process, communication, transportation, entertainment, food, travel, design, and more. But I suspect that these enormous repeated doublings, as Moore's law marches forward in the next decade and drives computing speeds up and prices down, will transform lives and industries in ways that we are only just starting to imagine.

Maybe Too Big To Fail, but Not Too Big to Suffer

To contact us Click HERE
Which financial institutions are "too big to fail"? According to a report from the international Financial Stability Board, a working group of governments and central banks that tries to facilitate international cooperation on these issues, here's the list as of November 2012.

 Ready for a nice bowl of acronym soup? This list is actually the "global systemically important banks," known as the G-SIBs, which are a subcategory of the "global systemically important financial institutions," or G-SIFIs.  Already finalized, as the Financial Stability Board (FSB) explains, are guidelines for the "domestic systemically important banks," the D-SIBs, which national governments are expected to implement by 2016. Meanwhile, the International Association of Insurance Supervisors (IAIS) has proposed a method of deciding who is a "global systemically important insurers," the G-SIIs. The Financial Stability Board (FSB) and the International Organization of Securities Commissions (IOSCO) are now working on a method to idenfity the systemically important non-bank non-insurance financial institutions (no acronym yet available).

Meanwhile, the Financial Stability Oversight Council (FSOC) within the U.S. Department of the Treasury is working on its own lists. In its 2012 annual report, it designated eight systemically important "financial market utilities"--that is, firms that are intimately involved in carrying out various financial transactions. Here's the list: the Clearing House Payments Company, CLS Bank International, the Chicago Mercantile Exchange, the Depository Trust Company, Fixed Income Clearing Corporation, ICE Clear Credit, National Securities Clearing Corporation, and the Options Clearing Corporation. (And I freely  admit that I have only a fuzzy idea of what several of those companies actually do.)

In addition, the Dodd-Frank legislation presumes that U.S. banks are systemically important if their h holdings exceed $50 billion in consolidated assets. David Luttrell, Harvey Rosenblum, and Jackson Thies explain these points, along with a nice overview of many broader issues, in their Staff paper on "Understanding the Risks Inherent in Shadow Banking: a Primer and Practical Lessons Learned," written for the Dallas Fed. For perspective, they offer a list of the largest U.S. bank holding companies, all of which comfortably exceed the $50 billion benchmark for consolidated assets.


This litany of who is "systemically important" feels disturbingly long, and it's only getting longer. But ultimately, it's a good thing to have such lists--at least if they lead to policy changes. Once you have admitted that a number of financial institutions are too big to fail, because their failure would lead to too great a disruption in financial markets, and once you have then made a commitment that government should not bail out such institutions, what policy prescription follows?

The proposal from the Financial Stability Board is that the G-SIBs (global systemically important banks, of course) should face a different set of regulatory rules. As the Dallas Fed economists explain, these could include "higher capital requirements, supervisory expectations for risk management functions, data aggregation capabilities, risk governance, and internal controls." There are two difficulties with this approach. First, it may not work. After all, a considerable regulatory apparatus in the U.S. did not prevent the financial crashes from 2007-2009. And second, it may work with undesired side effects. In particular, if there are heavy rules on one set of regulated financial institutions, then there will be a tendency for financial activities to flow to less-regulated financial institutions. "If regulation constrains commercial banks’ risk taking, many questionable assets may simply migrate to less-regulated entities."

I don't oppose regulating the SIFIs (that would be "systemically important financial institutions") more heavily. But it's important to be clear on the limits of this approach. After all, it's not that these institutions are big, but rather that they are so tightly interconnected with other institutions. As Luttrell, Rosenblum, and Thies explain: "TBTF [that would be "too big to fail] is not just about bigness; it also includes “too many to fail” and “too opaque to regulate."”

It seems to me that the key here is to remember that maybe some institutions are too big to fail, but they aren't too big to suffer! In particular, they aren't too big to have their top managers booted out--without bonuses. They aren't too big to have their shareholders wiped out, and the company handed over to bondholders--who are then likely to end up taking losses as well. One task of financial regulators should be to design and pre-plan an "orderly resolution" as they call it. The trick is to devise ways so that if these systemically important firms run into financial difficulties, the tasks and external obligations of certain large financial firms will not be much disrupted, for the sake of financial stability,but those who invest in those firms and who manage them will face costs.

The National Taxpayer Advocate Speaks Up

To contact us Click HERE
Yes,.there is a National Taxpayer's Advocate, and her name is Nina E. Olson. She has just released the National Taxpayer Advocate 2012 Annual Report to Congress. The report covers a lot of ground, each year looking at issues like taxpayer rights, taxpayer rights, identity theft, and other topics. Here, I'll just focus on one topic:  "The most serious problem facing taxpayers — and the IRS — is the complexity of the Internal Revenue Code." Here are some samples of the comments (footnotes omitted): 

• "According to a TAS analysis of IRS data, individuals and businesses spend about 6.1 billion hours a year complying with the filing requirements of the Internal Revenue Code. And that figure does not include the millions of additional hours that taxpayers must spend when they are required to respond to IRS notices or audits. If tax compliance were an industry, it would be one of the largest in the United States. To consume 6.1 billion hours, the “tax industry” requires the equivalent of more than three million full-time workers. ... Based on Bureau of Labor Statistics data on the hourly cost of an employee, TAS estimates that the costs of complying with the individual and income tax requirements for 2010 amounted to $168 billion — or a staggering 15 percent of aggregate income tax receipts."
• "According to a tally compiled by a leading publisher of tax information, there have been approximately 4,680 changes to the tax code since 2001, an average of more than one a day."
• "The tax code has grown so long that it has become challenging even to figure out how long it is. A search of the Code conducted using the “word count” feature in Microsoft Word turned up nearly four million words."
• "Individual taxpayers find return preparation so overwhelming that about 59 percent now pay preparers to do it for them. Among unincorporated business taxpayers, the figure rises to about 71 percent. An additional 30 percent of individual taxpayers use tax software to help them prepare their returns, with leading software packages costing $50 or more."
• "IRS data show that when taxpayers have a choice about reporting their income, tax compliance rates are remarkably low. ... [A]mong workers whose income is not subject to tax withholding, compliance rates plummet. An IRS study found that nonfarm sole proprietors report only 43 percent of their business income and unincorporated farming businesses report only 28 percent. Noncompliance cheats honest taxpayers, who indirectly pay more to make up the difference. According to the IRS’s most recent comprehensive estimate, the net tax gap stood at $385 billion in 2006, when there were 116 million households in the United States. This means that each household was effectively paying a “surtax” of some $3,300 to subsidize noncompliance by others."
• "From FY 2004 to FY 2012, the number of calls the IRS received from taxpayers on its Accounts Management phone lines increased from 71 million to 108 million, yet the number of calls answered by telephone assistors declined from 36 million to 31 million. The IRS has increased its ability to handle taxpayer calls using automation, but even so, the percentage of calls from taxpayers seeking to speak with a telephone assistor that the IRS answered dropped from 87 percent to 68 percent over the
  period. And among the callers who got through, the average time they spent waiting on hold increased from just over 2½ minutes in FY 2004 to nearly 17 minutes in FY 2012."
• "The IRS receives more than ten million letters from taxpayers each year responding to IRS adjustment notices. Comparing the final week of FY 2004 with the final week of FY 20102, the backlog of taxpayer correspondence in the tax adjustments inventory increased by 188 percent (from 357,151 to 1,028,539 pieces), and the percentage of taxpayer correspondence classified as “overage” jumped by 316 percent (from 11.5 percent to 47.8 percent)."
• "In 2012, TAS [the Taxpayer Advocate Service] conducted a statistically representative national survey of over 3,300 taxpayers who operate businesses as sole proprietors. Only 16 percent said they believe the tax laws are fair. Only 12 percent said they believe taxpayers pay their fair share of taxes."
• "To alleviate taxpayer burden and enhance public confidence in the integrity of the tax system, the National Taxpayer Advocate urges Congress to vastly simplify the tax code. In general, this means paring back the number of income exclusions, exemptions, deductions, and credits (generally known as “tax expenditures”). For fiscal year (FY) 2013, the Joint Committee on Taxation has projected that tax expenditures will come to about $1.09 trillion, while individual income tax revenue is projected to be about $1.36 trillion. This suggests that if Congress were to eliminate all tax expenditures, it could cut individual income tax rates by about 44 percent and still generate about the same amount of revenue."

Finally, I'm not the biggest fan of infographics,which often seem to me like overcrowded Powerpoint slides on steroids. But for those who like them, here's one from the National Taxpayer Advocate summarizing many of these points: 

Some Thoughts on James Buchanan

To contact us Click HERE
When I think of James Buchanan, who died last week, I remember hearing him answer a question at a conference in the late 1980s, not long after he had won the 1986 Nobel prize. A questioner with stars in his eyes asked how Buchanan had been able to write, not a few dozen journal articles, but a series of more than 30 book-length tomes like The Calculus of Consent during his career. (In fact, Buchanan's work was later collected into 20 volumes.) Buchanan paused for a moment, the very image of a courtly, soft-spoken, silver-haired Southern professor, and then gently drawled: "Apply butt to chair." Words to live by!

I had barely heard of Buchanan's work when he won the Nobel prize in 1986: it certainly didn't have much prominence in my undergraduate or graduate studies in economics. But Buchanan contributed an article, "Tax Reform as Political Choice," to the first issue of my own Journal of Economic Perspectives in Summer 1987. (As with all JEP articles from the first issue to the most recent, it is freely available on-line courtesy of the American Economic Association.)

At that time, Buchanan had just presented his Nobel lecture, "The Constitution of Economic Policy,"
which offers an overview of his arguments about how economists should do policy analysis. He quoted the great Swedish economist Knut Wicksell (1851-1926) who wrote things like: "[N]either the executive nor the legislative body, and even less the deciding majority in the latter, are in reality ... what the ruling theory tells us they should be. They are not pure organs of the community with no thought other than to promote the common weal. ... [M]embers of the representative body are, in the overwhelming majority of cases, precisely as interested in the general welfare as are their constituents, neither more nor less."

In a similar spirit, here's Buchanan in his own words: "Economists should cease proffering policy advice as if they were employed by a benevolent despot, and they should look to the structure within which political decisions are made. ... I called upon my fellow economists to postulate some model of the state, of politics, before proceeding to analyse the effects of alternative policy measures. I urged economists to look at the "constitution of economic polity," to examine the rules, the constraints within which political agents act. Like Wicksell, my purpose was ultimately normative rather than antiseptically scientific. I sought to make economic sense out of the relationship between the individual and the state before proceeding to advance policy nostrums."

The first issue of JEP back in 1987 had a symposium on the 1986 Tax Reform Act, and in his JEP paper, Buchanan applied his perspective to how one might think about the passage of a largely revenue-neutral tax reform that sought to reduce special deductions, credits, and deductions, and to use the revenue saved to reduce marginal tax rates.

Buchanan argued that the "political agents" who set tax policy in general prefer higher taxes, because they like to have control over more resources. They also like to offer tax breaks to special constituencies, who reward them with political support. But as the political agents offer tax breaks to special constituencies, they need to ratchet up tax rates on others--and as those rates get higher, it becomes harder to offer still more tax breaks. From the perspective of these kinds of political agents, the 1986 tax reform could thus be interpreted as a chance to wipe the slate clean: that is, start over again with lower tax rates and fewer tax breaks. But this just meant that the political agents could then again start their pattern of offering tax breaks and pushing up rates all over again.

"The 1986 broadening of the tax base by closing several established loopholes and shelters offers potential rents to those agents who can promise to renegotiate the package, piecemeal, in subsequent rounds of the tax game. The special interest lobbyists, whose clients suffered capital value losses in the 1986 exercise, may find their personal opportunities widened after 1986, as legislators seek out personal and private rents by offering to narrow the tax base again. In one fell swoop, the political agents may have created for themselves the potential for substantially increased rents. This rent-seeking hypothesis will clearly be tested by the fiscal politics of the post-1986 years. To the extent that agents do possess discretionary authority, the tax structure established in 1986 will not be left substantially in place for decades or even years."

Over time, Buchanan's prediction has held true: new tax breaks have been created, and marginal tax rates have been pushed back up to pay for them. Thus, one common proposal for dealing with the disturbing prospects of long-run budget deficits is to raise additional revenue by limiting many tax breaks, and then using some of the revenue for lower marginal tax rates and some for deficit reduction. For previous posts of mine about such proposals, see my February 2012 post on  "Tax Expenditures: A Way to End Budget Gridlock?" or my  August 2011 post on "Tax Expenditures: One Way Out of the Budget Morass?"

Of course, Buchanan's work is a continual reminder that even when economists are involved in drawing up some plans for a tax reform that would broaden the tax rates and lower the base, political agents will be the ones who actually draw up the plan, vote on it, and determine how much it changes each year. This insight about the centrality of politics is perhaps obvious to the average person, but many of us who focus on the economy can lose sight of it. Buchanan's work is one reason that in my own Principles of Economics textbook  (available here), I follow up all the chapters on what can go wrong in markets--monopoly, externalities, public goods, inequality, incomplete information, all the rest--with a chapter on political economy. As a teacher, I want to encourage students to be thoughtful skeptics of how markets work, but I also want them to be equally skeptical about the extent to which political agents will implement the sort of economically enlightened policies that will actually address the problems of markets. Buchanan's obituary in the New York Times is here. It closes with a great line from Buchanan, who once said: “I have faced a sometimes lonely and mostly losing battle of ideas for some 30 years now in efforts to bring academic economists’ opinions into line with those of the man on the street. ... My task has been to ‘uneducate’ the economists.”

Management in U.S. Manufacturing Firms

To contact us Click HERE
Everyone knows that good management matters, but in a research study, how do you define "good management"? For example, defining "good management" according to whether a company earns high profits would be circular logic; it would assume ("good management leads to higher profits") what needs to be proven. Instead, "good management" has to be defined in some at least moderately objective way, so it can then be compared to how firms perform.

Nicholas Bloom and John Van Reenen, along with various co-authors, have been leaders in academic work that bravely seeks to define "good management," surveys the management practices of firms,  and then looks at how firms with "good management" are performing. Much of their early work compared management practices across countries: for example, see their  article "Why Do Management Practices Differ across Firms and Countries?" in the Winter 2010 issue of my own Journal of Economic Perspectives.

But now the U.S. Census Bureau has for the first time used this approach as part of a survey of 30,000 U.S. manufacturing firms.  A group of authors include Bloom and van Reenen, together with
Erik Brynjolfsson, Lucia Foster, Ron Jarmin, and Itay Saporta-Eksten provide the first glimpse of these results in "Management in America," a discussion paper written for the Center for Economic Studies, which is based in the U.S. Census Bureau.

Before describing the results, it's useful to say a bit more about how "good management" is measured in this survey. They refer to it as "structured management," a term which is more neutral than  "good." But the basic concept is a list of 16 questions about management at the firm, divided into questions about monitoring, targets and incentives. They offer these examples:

"The monitoring section asked firms about their collection and use of information to monitor and improve the production process. For example, how frequently were performance indicators tracked at the establishment, with options ranging from “never” to “hourly or more frequently”. The targets section asked about the design, integration and realism of production targets. For example, what was the time-frame of production targets, ranging from “no production targets” to “combination of short-term and long-term production targets”. Finally, the incentives asked about non-managerial and managerial bonus, promotion and reassignment/dismissal practices. For example, how were managers promoted at the establishment, with answers ranging from “mainly on factors other than performance and ability, for example tenure or family connections” to “solely on performance and ability”? The full questionnaire is available on
http://bhs.econ.census.gov/bhs/mops/form.html." 

For each of the 16 questions, they arbitrarily say that the lowest possible answer is given a score of 0, the highest possible answer is given a score of 1, and in-between answers are given fractional totals. They then just sum up the total points to get a management score for each firm. A number of other questions collect information both about the firm and about the personal characteristics of the middle manager answering the survey. Those who want additional gory details on the survey process can head for the report. Here are some findings.

The obvious starting point is to figure out whether manufacturing firms that get higher scores for "structured management" are performing better in some way. They divide up the firms into tenths, or deciles, from lowest score to highest, and find that firms with higher scores do better on profits, productivity, output growth, research and development spending, and patents. Here are a few sample figures to give you a feeling for what these relationships look like. The relationships continue to hold after adjusting for all the obvious factors, like size of firm, industry type, and the like.




An additional conclusion is that there is a great deal of variation in structured management across firms. As they write: "[T]here is enormous dispersion of management practices across America: 18% of establishments adopt at least 75% of structured management practices for performance monitoring, targets and incentives; while 27% of establishments adopt less than 50% of these practices."

They also find that management scores tend to be growing from 2005 to 2010; that management scores are higher in larger firms; and that firms in the South and Midwest tend to have  higher management scores that firms in the West and Northeast.

This research is, if not quite in its infancy, still in its toddler-hood, and so it would be unwise to draw conclusions too aggressively. But some preliminary conclusions would be: 1) It's possible to come up with a survey tool that captures at least a significant portion of what is meant by "good management;" 2) This measure of management appears to be correlated with positive performance by firms on a number of dimensions; and 3) This measure of management differs quite substantially across firms.

But there are also difficulties at this stage. Good management, of course, isn't just a matter of responding in a certain way on survey questions, or announcing a certain set of policies. In real life, good management is also about implementing those policies, and even of having a culture within the company in which the implementation of those policies is widely practiced and accepted. It may be, for example, that companies which lack structured management have a range of other problems or issues not well-captured by the existing data. It may be that implementing more structured management in a company that isn't used to this approach will prove difficult process in many firms. I expect that future research will delve more deeply into these issues, and others. 

Checkerboard Puzzle, Moore's Law, and Growth Prospects

To contact us Click HERE
My father the mathematician first posed the checkerboard puzzle to me back in grade-school, perhaps on some rainy Saturday. His version of the story went something like this:

The jester performs a great deed, and the king asks him how he would like to be rewarded. The jester is aware that the king is a highly volatile individual, and if the jester asks for too much, the king might just kill him then and there. The jester also knows that the king views his promise as sacred, so if the king says "yes" to the jester's proposal, then the king will honor that promise. So in a way, the jester's problem is how to ask for a lot, but have the king at least initially think it's not very much, so that the king will give his consent.

So the jester clowns around a bit and then says: "Here's all I want. Take this checkerboard. On the first square, but one piece of gold. On the second square, two pieces. On the third square, four pieces, and on the fourth square, 8 pieces. Double the amount on each square until you reach the end of the checkerboard."

In the story, the king laughs at this comic proposal and says,  "Your great deed was so wonderful, I would have happily done much more than this! I grant your request!"

But of course, when the king starts hauling up gold pieces from the treasury, he will discover that 2 raised to the 63rd power, the final spot on the checkerboard requires about 9 quintillion gold pieces (that is, 9 followed by 18 zeros). 

I've had some sense of the power of exponential growth ever since.  But what I hadn't thought about is the interaction of Moore's Law and economic growth. Moore's Law is of course named for Gordon Moore, one of the founders of Intel, who noticed this pattern back in 1965. Back in the 1970s, he wrote a paper that contained the following graph showing how much it cost to produce a computer chip with a certain number of components. Here's his figure. Notice that the numbers of component on the horizontal axis and the cost figures on the vertical axis are both graphed as logarithm (specifically, each step up the axis is a change by a factor of 10). The key takeaway was that the number of transistors ("components") on an integrated circuit was doubling about every two years, making computing power much cheaper and faster.


Ever since I started reading up on Moore's law in the early 1980s, there have been predictions in the trade press that it will soon reach technological limits and come to and end. But Moore's law marches on: indeed, the research and innovation targets at Intel and other chip-makers are defined in terms of making sure that Moore's law continues to hold for at least awhile longer. Stephen Shankland offers a nice accessible overview of the current situation in an October 15, 2012, essay on CNET: ""Moore's Law: The rule that really matters in tech"  (The Gordon Moore graph above is copied from Shankland's essay.)

As Shankland writes: "To keep up with Moore's Law, engineers must keep shrinking the size of transistors. Intel, the leader in the race, currently uses a manufacturing process with 22-nanometer features. That's 22 billionths of a meter, or roughly a 4,000th the width of a human hair." He cites a variety of industry and research experts to the effect that Moore's law has at least another decade to run--and remember, a decade of doubling every two years means five more doublings!

It's hard to wrap one's mind around what it means to say that the power of microchipo technology will increase by a factor of 32 (doubling five times) in the next 10 years. A characteristically intriguing survey essay  from the January 10 issue of the Economist on the future of innovation uses the checkerboard analogy to think about the potential effects of Moore's law. Here's a comment from the Economist essay:

Ray Kurzweil, a pioneer of computer science and a devotee of exponential technological extrapolation, likes to talk of “the second half of the chess board”. There is an old fable in which a gullible king is tricked into paying an obligation in grains of rice, one on the first square of a chessboard, two on the second, four on the third, the payment doubling with every square. Along the first row, the obligation is minuscule. With half the chessboard covered, the king is out only about 100 tonnes of rice. But a square before reaching the end of the seventh row he has laid out 500m tonnes in total—the whole world’s annual rice production. He will have to put more or less the same amount again on the next square. And there will still be a row to go.

Erik Brynjolfsson and Andrew McAfee of MIT make use of this image in their e-book “Race Against the Machine”. By the measure known as Moore’s law, the ability to get calculations out of a piece of silicon doubles every 18 months. That growth rate will not last for ever; but other aspects of computation, such as the capacity of algorithms to handle data, are also growing exponentially. When such a capacity is low, that doubling does not matter. As soon as it matters at all, though, it can quickly start to matter a lot. On the second half of the chessboard not only has the cumulative effect of innovations become large, but each new iteration of innovation delivers a technological jolt as powerful as all previous rounds combined."

Now, it's of course true that doubling the capacity of computer chips doesn't translate in a direct way into a higher standard of living: there are many steps from one to the other. But my point here is to note that many of us (myself included) have been thinking about the changes in electronics technology a little too much like the king in the checkerboard story: that is, we think of something doubling a few times, even 10 or 20 times, and we know it's a big change, but it somehow seems within our range of comprehension.

But when something has already been doubling every 18 months or two years for a half-century--it is continuing to double!--the absolute size of each additional doubling is starting to get very large. I lack the imagination to conceive of what will be done with all this cheap computing power in terms of health care, education, industrial process, communication, transportation, entertainment, food, travel, design, and more. But I suspect that these enormous repeated doublings, as Moore's law marches forward in the next decade and drives computing speeds up and prices down, will transform lives and industries in ways that we are only just starting to imagine.

Why Has Health Information Technology Been Ineffective?

To contact us Click HERE

Health information technology is one ofthe methods often proposed to help rein in rising health care costs. The underlying storyis plausible: greater efficiency in dealing with the provision of care and the paperwork burden ofmedicine, and greater safety for patients as providers can be aware of pastmedical histories and ongoing treatments. However, at least so far, health information technologyhasn’t done much to reduce costs.  ArthurL. Kellermann and Spencer S. Jones ask “What It Will Take to Achieve theAs-Yet-Unfulfilled Promises of Health Information Technology” in the firstissue of Health Affairs for 2013 (pp. 63-68). (This journal is not freelyavailable on-line, but many academic readers will have access through librarysubscriptions.)
Back in 2005, a group of RAND researchers forecast thatrapid adoption of health information technology could save $81 billionannually. Kellermann and Jones essentially ask: Why hasn’t this vision come to pass?  Here are some of their answers (as usual,footnotes are omitted).
Health providers and patients have been slow to adopt information technology. “The most recent datasuggest that approximately 40 percent of physicians and 27 percent of hospitalsare using at least a “basic” electronic health record. … Uptake of health IT bypatients is even worse.”
Existing health information technology systems don't interconnect. “Aremodern health IT systems interconnected and interoperable? The answer to thisquestion, quite clearly, is no. The health IT systems that currently dominatethe market are not designed to talk to each other. … As a result, the currentgeneration of electronic health records function less as an “ATM cards,”allowing a patient or provider to access needed health information anywhere atany time, than as “frequent flyer cards” intended to enforce brand loyalty to aparticular health care system.”
Health care providers dislike the existing information technology systems.“Considering the theoretical benefits of health IT, it isremarkable how few fans it has among health care professionals. The lack ofenthusiasm might be attributed, in part, to the sobering results of studiesshowing that in many cases health IT has failed to deliver promised gains inproductivity and patient safety. An even more plausible cause is that few ITvendors make products that are easy to use. As a result, many doctors andnurses complain that health IT systems slow them down.”
Existing health information technology can raise costs. On this point,, the authors cite a New York Times article from last fall by Reed Abelson, Julie Creswell, and Griff Palmer. (Full disclosure, Reed Abelson was a friend of mine back in college days.) The NYT story reports: "[T]he move to electronic health records may be contributing to billions of dollars in higher costs for Medicare, private insurers and patients by making it easier for hospitals and physicians to bill more for their services, whether or not they provide additional care.Hospitals received $1 billion more in Medicare reimbursements in 2010 than they did five years earlier, at least in part by changing the billing codes they assign to patients in emergency rooms, according to a New York Times analysis of Medicare data from the American Hospital Directory. Regulators say physicians have changed the way they bill for office visits similarly, increasing their payments by billions of dollars as well."

Kellermann and Jones end with a plea that health information technology systems should be built o principles of interoperability, ease of use, patient-centeredness. I have no disagreement with the principles, but I would note that even within individual companies, it has often proven quite time-consuming and difficult to integrate information technology into operations in a full and productive way. Thus, it's no surprise to me that the health care industry has faced a number of stumbling blocks. I’ve heard anecdotal stories of doctorsspending inordinate amounts of time clicking through menus on some IT system,trying to figure out which boxes to check to best represent a diagnosis and acourse of care. I’ve heard that some doctors, as they master the system, findthat it becomes easier to bill for many separate small services that theywouldn’t have previously bothered to write up.
It seems that it should be possible for the big health care finance operation, both public and private, to get together and hammer out a basic flexible framework for health care information technology. But it doesn't seem to be happening.