Foreign Affairs on Outsourcing

[via Dan Gillmor] Foreign Affairs is one of the most influential publications in government circles. It has an article by Daniel Drezner on what has rapidly become a hot political issue in the US:

Critics charge that the information revolution (especially the Internet) has accelerated the decimation of U.S. manufacturing and facilitated the outsourcing of service-sector jobs once considered safe, from backroom call centers to high-level software programming. (This concern feeds into the suspicion that U.S. corporations are exploiting globalization to fatten profits at the expense of workers.) They are right that offshore outsourcing deserves attention and that some measures to assist affected workers are called for. But if their exaggerated alarmism succeeds in provoking protectionist responses from lawmakers, it will do far more harm than good, to the U.S. economy and to American workers.

Should Americans be concerned about the economic effects of outsourcing? Not particularly. Most of the numbers thrown around are vague, overhyped estimates. What hard data exist suggest that gross job losses due to offshore outsourcing have been minimal when compared to the size of the entire U.S. economy. The outsourcing phenomenon has shown that globalization can affect white-collar professions, heretofore immune to foreign competition, in the same way that it has affected manufacturing jobs for years. But Mankiw's statements on outsourcing are absolutely correct; the law of comparative advantage does not stop working just because 401(k) plans are involved. The creation of new jobs overseas will eventually lead to more jobs and higher incomes in the United States. Because the economy -- and especially job growth -- is sluggish at the moment, commentators are attempting to draw a connection between offshore outsourcing and high unemployment. But believing that offshore outsourcing causes unemployment is the economic equivalent of believing that the sun revolves around the earth: intuitively compelling but clearly wrong.

Should Americans be concerned about the political backlash to outsourcing? Absolutely. Anecdotes of workers affected by outsourcing are politically powerful, and demands for government protection always increase during economic slowdowns. The short-term political appeal of protectionism is undeniable. Scapegoating foreigners for domestic business cycles is smart politics, and protecting domestic markets gives leaders the appearance of taking direct, decisive action on the economy.

Protectionism would not solve the U.S. economy's employment problems, although it would succeed in providing massive subsidies to well-organized interest groups. In open markets, greater competition spurs the reallocation of labor and capital to more profitable sectors of the economy. The benefits of such free trade -- to both consumers and producers -- are significant. Cushioning this process for displaced workers makes sense. Resorting to protectionism to halt the process, however, is a recipe for decline. An open economy leads to concentrated costs (and diffuse benefits) in the short term and significant benefits in the long term. Protectionism generates pain in both the short term and the long term.

Two more stories on outsourcing:

WSJ: "U.S. companies sending computer-systems work abroad yielded higher productivity that actually boosted domestic employment by 90,000 across the economy last year, according to an industry-sponsored study...The study's premise is that U.S. companies' use of foreign workers lowers costs, increases labor productivity and produces income that companies can use to expand both in the U.S. and abroad. It was commissioned by the Information Technology Association of America, an industry membership and lobbying group, which hired the economics consulting firm Global Insight Inc. of Lexington, Mass...The study claims that twice the number of U.S. jobs are created than displaced, producing wage increases in various sectors. The report takes a rather narrow focus, tracking the outsourcing of computer-services jobs, but not other work increasingly being done abroad such as manufacturing, call centers or medical X-ray reading."

News.com: "The U.S. technology industry's demand for offshore services is apparently beginning to drive up pay rates in India, raising questions about the long-term benefits of outsourcing work to that country...India's wage inflation, which approached an estimated 14 percent last year, is a natural byproduct of a classic supply-and-demand scenario."

A historical perspective on outsourcing is provided by another WSJ story:

Losing skilled jobs to low-wage foreign competition is as old as the Industrial Revolution. In the 1830s, the British textile industry became so efficient that Indian cloth makers couldn't compete. The work was outsourced to England, with disastrous consequences for Indian workers. "The misery hardly finds parallel in the history of commerce," India's governor general, William Bentinck, wrote to his superiors in London in 1834.

As Americans grapple with the fallout of shipping hundreds of thousands of jobs overseas, history echoes with many similar episodes -- and lessons. Trade and technology can boost living standards for many people, by creating lower-priced goods. But those same forces can destroy skilled jobs that workers thought never would be threatened.

Competition from foreign labor hurt huge classes of American workers in the 19th century but eventually helped ease wage disparities between nations. And during these upheavals, history shows that politics can arrest what seems like unstoppable technological progress.

Here are four lessons from history that help illuminate today's debate:
- Even high-skilled, good-paying jobs are vulnerable.
- Trade liberalization often works with technology to undermine powerful interests.
- Domestic workers are always vulnerable to competition from foreigners willing to work for less.
- Politics can slow down the transforming effects of new technology.

Tim Bray visits OpenOffice

OpenOffice is what I use on the desktop. Tim Bray just joined Sun, and one whose blog I find fascinating reading. So, when there is a post about the combo, I have to read and blog it!

What’s In the Package: A word processor, a slide show maker, a spreadsheet, a vector-graphics package, a database client; more or less what the competition has, minus a standalone database, plus better graphics.

XML! The way that these guys store the data is massively, fiendishly, outrageously clever. They have their own XML tag set, which includes (in one namespace) all the basic word-processing, spreadsheet, and slide-show machinery. Then, for graphics they use SVG, for styles they use XSL-FO, for links they use XLink... you get the picture, they’ve invented the absolute minimum possible.

As if this wasn’t clever enough, they wrap up documents in a zipfile with a manifest and a MIME type and separate chunks of XML for the data and metadata and styles and and manifest and so on. So the size is moderate, it loads fast, and it’s all in a single handy blob.

An idea from Tim: "It turns out that OpenOffice already comes with a doohickey that will produce an XHTML approximation of most documents (Lauren tells me it’s shaky on tables); plus it’s got a nice HTTP library and APIs out the wazoo. Can you see what I’m thinking? There’s no reason this sucker shouldn’t have a 'Blog this' button that XHTML-i-fies whatever you’re typing, lets you preview, and then lets you ship it out via one of the existing blogging APIs or the Atom API."

ODP, RSS and OPML

Dave Winer writes:

The Open Directory Project is going to do something with RSS, not sure what, but it's a good sign, if only just a start. Here's what I would like them to do.

1. Associate a feed with a level of the hierarchy, so someone can subscribe to a category, and anything that appeared in that category would show up in the reader's aggregator as new.
1a. Associate a feed with a level in the other direction, so that news can be routed to a category in the directory. So, to the left, you'd see the stuff that doesn't change often, and in a box to the right is the new stuff.

2. Let an author maintain a whole level of the directory with RSS.

3. What about more than one level? We thought of that too, it's called OPML.

4. After adopting RSS and OPML, implement inclusion, meaning you can point to an OPML file anywhere a node can appear and the content of that OPML is included in the directory as if it were part of the directory.

5. From there, the whole thing will be unbundled, let the search engines understand an OPML file and display the as Yahoo-like directories.

I had written about how some of these elements taken together could be used to construct the Memex.

Convergence 2.0

Om Malik points to a post by Christian Lindholm on some lessons from Cebit:

1. It is getting possible to build the life recorder: Integrating voice, image, video and text is becoming possible in an uncompromised form factor known to the broad audience as a mobile phone.

2. The Laptop is becoming a personal communications tool: The integration of Wi-Fi totally changes to nature of a laptop. Laptops are now powerful enough and small enough to really become communications tools.

3. The Home server, rebranded as an entertainment center is the future of the Home PC: Intel was showing off the "Kessler" concept which integrated PC, Wi-Fi, RAID and lots of other technology I still do not understand, but it left a permanent impression on me.

Om calls it Convergence 2.0 and Marc Canter calls it Digital Lifestyle Aggregation. It is also part of the vision of Dana Blankenhorn's Always-On World.

Social Software

Jon Udell writes:

Computer-mediated communication is the lifeblood of social software. When we use e-mail, instant messaging, Weblogs, and wikis, we’re potentially free to interact with anyone, anywhere, anytime. But there’s a trade off. Our social protocols map poorly to TCP/IP. Whether the goal is to help individuals create and share knowledge or to enrich the relationship networks that support sales, collaboration, and recruiting, the various kinds of enterprise social software aim to restore some of the context that’s lost when we move our interaction into the virtual realm.

Whatever the mode of communication, the primary goal, says Adam Hertz, VP of technology strategy at Ofoto (and a Socialtext user), is to create group memory. Chris Nuzum, CTO and co-founder of Traction Software (infoworld.com/1054), echoes that theme. Traction describes its offering as “enterprise Weblog software,” but Nuzum says that a typical Traction project is more of a group effort than an individual journal. As such, a lot of the social interaction that would otherwise occur in e-mail moves into the comments and discussions attached to the project.

Building group memory and team awareness has always been the goal of KM (knowledge management), of course. “But most people,” Nuzum says, “have never had the benefit of mechanized institutional memory.” One reason for this limitation is that KM systems have tended to ask people to dump knowledge into databases without regard for social incentives, habits, or consequences. These are central concerns for social software in all its various forms.

TECH TALK: As India Develops: ICT (Part 2)

What India needs is an affordable computing platform to build out the digital infrastructure across its enterprises, homes, education institutions and government. So far, the high cost of the computers and even higher cost of software (relative to income levels) has hobbled adoption of technology in India. Luckily, the elements to construct solutions at price points which are 70-90% lower without sacrificing performance are now available.

Thin clients are computers with limited local processing and storage, which are both centralised on servers. The client should be able run an OS (Linux ,for example) along with VNC (virtual network computer). VNC is a remote display protocol, much like RDC from Microsoft and ICA from Citrix. Using this approach, thin clients can be, theoretically, assembled for less than Rs 2,250 (USD 50) in component costs, excluding the display. A refurbished 14-inch monitor will cost under Rs 2,000 (while a new one will cost about Rs 3,500). TV can be a possible display option also, though one will have to sacrifice the viewing quality. Thus, it is possible to put together the user desktop for no more than Rs 5,000.

Server-centric computing is just like the mainframe and minicomputers of the past. Terminals managed the local input/output while all the computing and data storage took place on the server. A return to a similar model is essential if one has to dramatically bring down the cost of computing. In today’s world, there are two things working in our favour: networks (local and wide-area) have become fast enough to enable data to be transmitted rapidly across from the server to the client, and Moore’s Law provides for huge server processing capabilities at much lower price points. So, it should be possible to centralise computing at a cost of no more than Rs 2,500 per user (for a minimum of 10 users on the LAN) or a fraction of that if being offered as a hosted service by broadband operators to homes.

Open-source software is the third leg of the affordable computing. For almost every commercial software there is a free, open-source equivalent. For the desktop, Linux compares well with Microsoft Windows, OpenOffice with Microsoft Office, Ximian Evolution with Microsoft Outlook, Mozilla with Internet Explorer. For the server infrastructure, it is possible to put together the complete infrastructure of a mail server, proxy server, firewall, anti-virus software, anti-spam, database server and file/print server software with open-source components. In addition, there are thousands of other open-source applications available for specific needs.

Taken together, thin clients, server-centric computing and open-source software offer an excellent platform on which to build an alternate environment which can dramatically reduce the cost of computing for the next markets. By leveraging a handset-like business model, service operators (think of them as “tech utilities”) can start offering hardware, software, networking, connectivity and support for as little as Rs 500-600 per month, bringing the cost of computing down to that of a cellphone today. This is what is needed for building out India’s digital infrastructure.

Tomorrow: ICT (Part 3)

Related Entries:  [All]