As I write this, the Dow index fell 500 points in one day, Lehman Brothers has just been bought by Barclays of London after the likelihood of bankruptcy and Bank of America is buying up the distressed assets of Merrill Lynch after doing a similar service for Countrywide Mortgage earlier this year. The Federal Reserve has effectively "bought" Freddie Mac and Fannie Mae in order to keep them propped up - and rumors have insurance giant AIG next in line for being put into conservatorship.
Meanwhile, Tropical Storm Ike, even played out as a hurricane, continues to dump massive amounts of wind and water over much of the midwest as far north as Montreal, flooding parched corn and soybean fields with far too much water, likely ruining harvests that were already marginal at a time when demand for these two staples is still extraordinarily tight. Add into that the devastation caused by Ike both to the economies of southern Texas and the stunning jump in gas prices all across North America as a consequence of Ike (here in Victoria, BC prices jumped from $1.37 a liter to $1.50 overnight, a rise of fifty cents a gallon, despite Galveston being 2000 miles away), and this weekend has become quite literally an example of a Perfect Storm.
O'Reilly's focus has long been on programming issues (or programmer issues) and that focus remains very much in place. However, it is worth understanding how the grief playing out on Wall Street will have a very significant impact upon the IT industry within the next four to six months, even despite the fact that up until now the contagion seems largely to have remained contained in the financial sector.
A System in Collapse
What is occurring now is a series of events that has not really happened in more than eighty years, and in some respects has never happened at the magnitude that's occurring now. For a number of reasons - non-existent oversight, loose economic policies by both Republican and Democratic administrations, the decision to break down the barriers between banks, investment houses and insurance companies, predatory (and likely criminal) lending practices (and a public that didn't do its homework) and the ability to use the Internet to turn everyone into speculators that swamped the system with negative feedback loops - the financial sector is now in catastrophic collapse.
This is not hyperbole so much as a description from systems theory. Most banking works upon the central idea that when a bank provides a loan, the loan is predicated upon the idea that the risk of providing the loan is compensated by interest payments, so that in the end, the bank makes a certain amount of money for taking the risk of not getting paid back at all. It also makes money originating the loan in the first place, which is generally a non-risk event.
In order for this to happen, however, the bank also needs to make sure that in the event that a loan goes bad (isn't paid back or is paid back slower than was agreed upon) that they have a certain degree of protection that the money that they're out will not impact upon other customers of the bank. This is the bank's reserves. Such reserves were, for a long time, kept at roughly 10% of the total loans outstanding for the bank. This means that if they had ten accounts, they could effectively handle one of those going bankrupt, assuming all the loans were of equal size.
Put another way, a bank could effectively leverage a million dollars of reserves into nine million dollars of loans. This is why when the Fed commits $100 billion to banks, it effectively creates $1 trillion dollars in effective loans, which is why such Fed money is often called "super-money".
Banks have been trying to chip away at the 10% cap since the 1960s, since it limited the number of loans that they could originate. If this were a political blog I would go into the effects that this initiative has had on politics, but it isn't and I won't. The fractional reserve system did begin to score big in the mid-1990s as the reserve fraction was systematically lowered, so that it is now roughly 1% (the exact value depends upon a lot of factors). This means that in essence $10 trillion dollars can be secured by about $100 billion in reserves.
Of course, remember that this fraction effectively is a measure of perceived risk. At 99:1, this meant that if more than one loan "went bad" out of 100, then the bank would have to borrow from other banks (or from the "superbank" - the Federal Reserve (a.k.a, the Fed). Most of this happened in the background - one or two banks at a time might have trouble with a problematic loan for a short term, but they could always count on other banks helping them making up the difference in exchange for doing the same for the other banks at some point.
Investment banks are a little different from traditional banks in that their primary mission was to return yield to their investors. Investment banks assumed higher risk as a consequence, of course (that's essentially what a yield is - the cost premium for risk), but they could often get much better marginal returns than more traditional banks. Lehman Brothers, Merrill Lynch and Bear Stearns were all investment banks first and foremost.
The Gramm-Leach-Bliley act of 1999 was instrumental for the banks in that it let regular banks get into both the investment and mortgage business. This has led to the consolidation of a number of "full-service" banks that can loan out commercial and home mortgages, provide investment services, even sell insurance - and has also blurred the distinction between these various types of services. This act caused a flurry of consolidation, eliminating many of the small "traditional" local banks in favor of branches of banking behemoths - and not coincidentally, greatly increasing the potential that a catastrophic risk event for a bank would have huge negative repercussions for the industry.
After 9/11 (and perhaps using it as an excuse) the Federal Reserve also dropped interest rates that it charged to these banking behemoths to as low as 1%, and effectively held the rates there. In essence, this meant that, as inflation was running "officially" at about 2.5% over the last decade that the Fed was effectively giving the banks super-money at no risk. Combine this with nearly eliminating the fractional reserve, and it was the equivalent of goosing the economy with "free" money (I'll get back to that point in a second, however).
A second change that took place in 2003 was the Bush administration's "Ownership Society", which in essence made it possible for banks and mortgage companies that originated mortgage loans to sell them to other parties at little cost or penalty. Consider the implications of this - the return on a mortgage is fairly minimal, nominally between 3% to 5% at the height of this silliness, while the risk of default is fairly high. If a bank could in essence originate a mortgage and sell it if they felt the risk of default was too high, then they make money on the transaction fees while keeping the risk to their primary assets fairly minimal.
This of course also meant that it was in the best interest of banks to sell expensive mortgages to people to maximize their fees, even if it was obvious that the mortgage burden would be crushing to the lendee. Housing prices consequently went up (there's some evidence to indicate strong collusion between lenders and real-estate appraisers here that is still just beginning to shake out) as demand outstripped supply. After all, who wouldn't want to get into a new house at such low "introductory" rates, especially if you had problem credit to begin with.
Similar processes were occurring elsewhere, of course - finance expensive cars at low payments and low rates (but with high origination fees) and watch the happy owner drive off the lot, even if they probably couldn't have afforded the car in the first place. Make sure credit card applications are everywhere so that people could put themselves in debt even faster, because ultimately the originator of the all of these loans no longer had the obligation of assuming the risk of default.
Of course, to make this debt palatable, the purchaser of the debt also assumes the risk, so many banks would try to bundle up "good debt" along with bad so that purchasers would be more comfortable in owning this. Yet with the ability to disassemble and resell such debt, what ended up happening was that the bank fractioning system began to resemble an oil fractioning system - the profits got skimmed out at each stage of the process, while what remained became increasingly more risky - and increasingly difficult to identify as such.
The problem with this process is that it eventually becomes a game of 'hot potato' (though radioactive potato may be closer to the mark). At first, much of this debt was sold as "investment grade vehicles" to other banks, foreign governments and municipalities (in order for them to finance bonds and other initiatives). Yet as the Fed started raising interest rates (and increasing the pressure on people who had bought on houses they either couldn't really afford or because of adjustable rate mortgages (ARMs) causing the mortgage to increase steadily - or in some cases, jump dramatically - each month) more and more of them began to default - and people who held these vehicles began to get nervous. The banks found it harder to sell them, and were forced to hold onto them, though in many cases "off the books" - that is to say, they didn't have to account for them against their existing reserves.
What's more, more and more of these found their way onto the books of two quasi-governmental corporations intended to make home ownership possible for most Americans - the Federal National Mortgage Association and the Federal Home Loan Mortgage Corporation (better known as Fannie Mae and Freddie Mac), as these were essentially the guarantors of last resort for the banks to get the money for the mortgages in the first place.
In early 2007, the mortgage crisis began to hit as people began to default on the lowest quality loans, and these houses went into foreclosure. Banks do not like for houses to go into foreclosure - it makes them liable (somewhat) for property taxes, and they also become responsible for managing the properties. They are in fact lousy investment vehicles unless someone actually lives in them. When the property market was hot they could of course resell the property, but as prices reached a level above what most people could pay and as other factors, such as inflation due to gasoline and food costs along with static wages ate into people's after tax income, it became harder to sell these foreclosed properties, and they also began to sit, still unrealized as obligations, on the bank's books.
Countrywide Mortgage failed in 2007 because it had too much property and too many bad debts on its books (while they were still originating mortgages, they were finding it more and more difficult to sell them). This in turn served to wake people up to the problems in the industry, and for investors to start wondering what they had invested in. People stopped buying these particular vehicles fairly rapidly thereafter, and by October 2007 the large investment banks were in a position where they were left holding huge amounts of toxic mortgages that no one would touch, and investors were letting them know that they wouldn't purchase any securities until the extent of what that debt was became known.
The problem with that is that with no buyers, this essentially meant that these banks now had to effectively move these assets "onto" their books, which in turn meant that it counted against their reserves. Since in many cases these assets had been given as their "putative" values to bolster their reserves, this meant that overnight banks that had appeared to be financially solvent no longer had the financial reserves necessary to sustain the loans that they had outstanding, and as a consequence were forced to borrow these funds or attract new investors at a time when investors were eyeing them warily.
What this has meant in practice is that the banks have reached a vicious cycle - they can't loan money if they don't have the fractional reserve, and consequently this has caused the credit market to stall - people or companies cannot get loans. Investment into the banks have slowed to a trickle as more and more banks fail. What's worse, the loans outstanding that are going bad are consequently climbing as the economy gets starved of (credit-based) oxygen, which in turn is forcing more write-downs of what had been seen as previously good quality debt.
Enter the Fed. The Federal Reserve is not in fact a governmental agency. Instead, it is a quasi-governmental organization made up by the banking industry, though the President has the power to appoint governors. The Fed in essence is able to ask the Department of the Treasury for loans based upon existing or future taxpayer revenues. This means that when the Fed provides a $30 billion loan to a bank, what is providing the collateral for that loan are your taxes. In most cases, the bank does have to pay the loan back within a window of time, from overnight to 90 days, with a certain interest rate (one of these is the "Feds Fund rate" that is usually quoted in the evening news).
As these banks continue to fail, the Fed is now having to step in and essentially become guarantors of all of these toxic loans - the US government has essentially assumed the entire risk of the credit markets defaulting. Of course, as events of the last week have shown, you can only become a guarantor of a loan if people believe that you are capable of paying that loan back.
If people believe that you might have trouble paying the loan bank in a timely fashion, they require a higher premium, meaning that you have to pay more of your promissory notes (i.e., dollars in this case) for goods - and investors require a higher rate of return (i.e., interest rate) on their investments. If they begin to doubt your ability to pay at all, then your notes become increasingly worthless ... which causes the dollar to fall in relationship to other currencies.
The Economy and IT
This saga is still unfolding, but already its impacts are being felt. People, like their government, have been living on borrowed credit for some time as wages have failed to keep pace with inflation (which is the aforementioned fall of the dollar as people lose faith in your ability to repay). In addition to letting people purchase new homes, the loose credit of previous years also made it possible for people to take out second mortgages against the collateral of their house, at first for luxuries, then increasingly for necessities as the imbalance between wages and prices grew.
Near the top of the tech boom, average programmer salaries were around $70,000 for someone with 6-10 years of experience. In the decade since then, the dollar has lost approximate 48% of its value. This means that the same salary today would be worth only about $35,000 dollars ten years ago. Put another way, you would need to earn $135,000 today just to keep up with inflation.
A loss of 48% of value seems extreme, but is has to do with fact that inflation is like interest rates - it compounds. Thus, a 2.5% inflation rate means that the actual value of the currency drops to 97.5% of its value. 2.5% of that new value takes you to .975 * .975 or 95%. At no point in the last decade has the inflation rate been below 2%, and currently (officially) it's at around 5.5% - taking all of these numbers together you get a 48% drop.
One thing to keep in mind, however, is that in the late 1990s, the inflation computations were changed to remove oil and food from the measure, as well as using such constructs as rent equivalents (as housing prices were beginning to grow dramatically) as well as the use of hedonic deflators - arguing that a computer that was twice as fast and powerful then a previous model in that line was more efficient, thus it was only half as expensive and similar equivocations. If you go back to the previous measure of inflation, the drop is closer to 65-75%.
Moreover, wages have generally remained stagnant, and now computer hardware and software manufacturers are beginning to trim staff. HP announced today (after hours) that it would cut 24,600 jobs from their payroll, after similar announcements through this week. As demand for goods and services has dropped off due to an increasingly constrained consumer, the ability of companies to pay for salaries drops off accordingly.
This round will probably hurt companies far more than it did in the tech recession - at that time, there were too many people in IT and the reduction was a fairly typical realignment of supply and demand. This time around, there is both a perceived (and to a certain extent real) skills shortage in the industry and as the dollar continues to fall (despite its recent - likely manufactured - rise) outsourcing is becoming cost prohibitive. This is probably one of the reasons why IT has seemed immune to the present troubles, but that won't last - it just means that the cuts are likely to come a little later for IT professionals than they would for sales or marketing.
From a business standpoint, the timing of this couldn't have come at a worse time either. Computer systems have a lifespan of about 6-8 years, and the last major investment in infrastructure occurred around 2000. This means that many businesses are now facing applications and hardware that are reaching the end of their useful lifespan, but as most capital infrastructure upgrades typically assumes taking out loans rather than pulling funds out of operating cash (which is likely getting strained as well) this will mean that future programming and hardware initiatives are now being cut.
Programming and engineering salaries similarly tend to be at the upper end of the scale for non-managerial positions, which means that as the recession that is occurring due to the credit crunch lengthens, the degree to which such professionals keep their jobs will be as much due to the strength of the CTO in the organization as any other factor.
Additionally, it is also likely that contract agencies will also be hit harder, as the typical contract programmer is usually brought in as a project is in the ramp up phase, in part because if such a project does not succeed the programmer can be let go less expensively than letting go a full time employee. WIth fewer projects, this translates in the long run to fewer contracts and likely downward pressures on wages as a consequence. The one exception to this is the situation where a contractor is brought on to perform maintenance work on existing systems, which typically means 2 to 4 month contracting assignments.
Similarly, the economy will impact stand-alone proprietary system development companies far more than it will open source and services oriented developers. Companies such as Microsoft will likely be fairly hard hit both because the number of computer systems sold (and hence software licenses for Microsoft Windows sold through their OEM channels) will continue to drop at both the consumer and enterprise level, while they will be competing not only with open source (and by extension free) software packages but also will be contending with their own legacy applications.
Given that Microsoft has not managed to get the services foothold in the office space that they so desperately have been trying to achieve, it may very well be that it will be too late now for them to get in at all - this musical chairs game may have ended with one too few chairs. This does of course lend new urgency to incoming Chief Architect Ray Ozzie to develop an effective online strategy now, and I suspect that it may in the end ultimately require making some hard choices about how far they want to carry the Windows and Office legacies in a business environment that will become very hostile soon.
This situation does not necessarily play much better for Google. At the end of the day, Google's bread and butter is still advertising, which typically ends up tanking in the face of a recession, especially one as broad as this one is boding to be. While it is likely that Google will continue to spend on its own internal R&D for some time, it is likely that there will also be considerable belt tightening there. The same issues also hold (perhaps more so) for companies such as AOL and Yahoo, although AOL has managed to strengthen its own advertising base considerably after Google's 5% investment in them last year.
I bring Google and Microsoft up not to start a detailed company analysis but to highlight the challenges facing the "top of the food chain". Overall, conditions are going to deteriorate for large and mid-sized software development companies. I suspect that even those that are depending upon government contracts are going to be challenged moving forward - by most estimates the Treasury Department has effectively committed anywhere from $400 billion to $1.5 trillion in funds just in the last couple of months and a conservatorship deal for insurance company AIG is rumored to be in the works as well (likely to be as large as the Freddie Mac/Fannie Mae "conservatorship").
What this means in simplest terms is that moving into next year, the accounting confusion alone is likely to significantly reduce the ability of the U.S. government to spend on much of anything outside of already extent line items, and many of those will be postponed or cancelled outright if they aren't absolutely critical.
Open Source's Shining Moment?
For all the talk of doom and gloom, the picture actually doesn't look quite that bad for small to small/medium sized companies (though it still isn't good). Most smaller tech companies are privately funded, either by VC rounds or the funds of the founders. Additionally, many of these companies typically tend to have comparatively low levels of overhead, especially for web application development shops, for those companies that have a part of their workforce working online or for those companies that are providing online services.
Indeed, it is quite likely that as applications become more distributed and virtualized, one consequence of this will be the strengthening of a trend to incorporate such services into existing applications rather than developing these capabilities in-house. This is ultimately what "mash-ups" are all about, but as a phenomenon mash-ups have remained largely a consumerist rather than an enterprise phenomenon. Yet by going to such a modular approach, it is likely that a lot of organizations looking to push their existing systems a little farther can integrate these services much more inexpensively than it would cost to build such systems in house.
While in the past few years, open source's appeals had more to do with the freedom from intellectual property disputes as much as licensing fees, the quality of such tools is also reaching a level where the cost argument (even with ROI) makes FOSS software far more appealing to the enterprise market. This argument was fairly weak when people considered migration costs and difficulties, but the effective credit freeze is likely to make this particular argument much more compelling.
This transition, coupled with the services transition mentioned early, will likely mean that more manpower will be devoted (and funds reallocated) to creating and improving upon open source projects as a cost effective alternative to in-house development or commercial software acquisition.
At the level of developers and IT professionals, the drill is essentially the same one that should have been learnt by programmers through the tech recession of 2001-2003.
- Stay liquid - the economy is going through some severe gyrations, and guarantees of the safety of bank accounts and investment vehicles should be taken with a healthy dose of skepticism. Make sure that you have cash on hand to cover emergency expenses.
- If you can, diversify your clients as much as possible, and if you work for a company, make sure that your resume is up to date, readily accessible on the web and that you are prepared in the event of that company suddenly having a spontaneous farewell party.
- Try to build up several months savings (admittedly tough in this day and age).
- Get involved in an open source project or two. Regardless of the platform that you're on, there's typically something out there that needs to have help, and as people start to use the project for their own need, the developers are typically the first ones to see work inquiries.
- Start a blog, especially a technical one - the more that you're seen as the go-to guy on the web in a certain area, the more likely you'll have clients coming to you.
- Stay away from trolling the job boards, or consider automating the search by retrieving feeds. The upcoming dislocation is already throwing a lot of people out of work, and chances are good that the job-bots on the employers sites will be filtering out a huge amount of resumes unless they happen to fit the exact right buzz-words.
- Network! Facebook, LinkedIn, Spock, whatever best fits your particular niche. Do this as insurance before you lose your job.
Estimates of the problems facing the economy vary dramatically. During the Savings and Loan crisis in the late 1980s, approximately 170 banks failed. It's possible that the situation right now is stabilizing. It's also possible (indeed, likely) that the financial tremors facing the largest banks will start working their way down to smaller banks, and the number of potential bank failures could be several times that. Either way, programmers and IT professionals should keep a wary eye on the financial storms that are brewing right now.
Kurt Cagle is Online Editor for O'Reilly media.