I first wrote about THSR in Puente AVE, an analysis of recent HSR traffic numbers and economics. This is an extended and updated version. There is a lot of shared information with the Wikipedia article because... it was me who put it in there too.

The incomplete completion

To uncover the problems at the root, let me start at the surface before digging deeper (and thus go back in time): the first thing to consider is that, despite an announced delay of almost a year, THSR opened half-finished.

The terminus in downtown Taipei (the capital) did not open for months after the line opened. Most wayside stations lacked urban mass transit connections (bus connections had to be established, urban rail connections are in construction now). An extension to the eastern suburbs of sprawling Taipei on one end, another from the current terminus in the suburbs of Kaohsiung to downtown at the other end (grey on map below), and three wayside stations (yellow on map) were delayed to a later date (some are in construction now).

Route map from Wikipedia. The line is currently 335.5 km long, non-stop trains travelling at 300 km/h cover it in 96–120 minutes (depending on the number of stops)

Trains themselves started with a reduced schedule, and with hefty ticket prices (standard adult ticket for a full-length one-way trip is NT$1,490 [c. €36.5 at current exchange rates], compare to NT$845 on conventional trains and NT$400-500 on buses), which the operator was forced to augment with discounts later.

Now, on one hand, under these circumstances, not meeting initial ridership predictions (originally: 180,000 a day, later reduced to 140,000) is no surprise at all, and the problem is a temporary one. On the other hand, these delays had their reasons.

Construction complications

Taiwan's west coast is a densely populated strip, comparable Japan's east coast. Such an area is ideally suited for a HSR line with traffic demand well beyond anything in Europe, but also calls for lots of expensive superstructure: bored tunnels across mountains, cut-and-cover tunnels or elevated sections to protect urban areas, and long elevated sections across swamps and agricultural land with unstable ground.

A THSRC 700T on a curved long viaduct near Shetou, south of Taichung station, on 17 April 2010. Photo by Johnson Wang from Flickr under CC-by-nc-sa 2.0

High infrastructure costs also bring the risk of high cost overruns. In THSR's case, cost overruns were largely the consequence of a switch of suppliers at a point in time when planning was already well-advanced.

In the competition to supply THSR, Alstom (one of the France-based companies building the TGV) and Siemens (one of the Germany-based companies building the ICE) teamed up in the Eurotrain consortium, which won the status of preferred bidder. Then they worked out infrastructure specifications, and demonstrated the hybrid train they offered: the more powerful tractor heads of two ICE 2 sets were attached to the lighter middle cars of a TGV Duplex set.

The Eurotrain at München-Laim. In a demonstration run on 4 May 1998, it easily achieved 316 km/h. Photo from André Werske's HSR site

The decision to ditch the preferred bidders and award the rival Japanese consortium instead was widely rumoured to have been political. It led to an epic political, media and court battle, ending in damage payments to Eurotrain. (As mentioned in Globalisation catches up with rail industry?, this heralded the coming of bidding duels until the last legal instance as regular competition practice in the industry.)

However, in the meantime, overseeing company Taiwan High Speed Rail Corporation (THSRC) stuck to its original specifications, forcing changes relative to the Japanese offer that weren't cost-neutral, for example:

• tunnels were bored with the large diameters meeting European safety regulations, rather than barely wider than the trains as on Shinkansen lines in Japan;

  

  A THSRC 700T enters a tunnel near Tongsiao, north of Taichung, on a pre-opening test run on 28 April 2006. Photo by user jiadoldol from Flickr under CC-by-nc-nd 2.0

• a German maker had to be contracted to supply special swing-nose slab track switches (fit for passage at high speed), because operation rules foresaw the possibility of changing tracks as on European mainlines, rather than the Shinkansen norm of strictly single-direction track usage (BTW, the switches malfunction);

  

  Commissioning of a RHEDA 2000® switch, made by RAIL.ONE, at the north end of Taichung station on 15 September 2005. The swing-nose is well visible along the centreline of the switch. Photo from Chan Chun Construction Co.

• THSRC retained a pool of train drivers from Germany and France, who were contracted to operate trains until domestic drivers are sufficiently trained.

Eventually, project costs grew to $18 billion. However, the real problem was not the (elevated) fixed costs themselves.

Let there be private finance

The horror story that unfolds when one looks into THSRC's finances has much to do with the current economic paradigm: the assumption, shared by governments and economists alike, that private risk-taking is always preferable to government control in the economy. Thus, in the nineties, the Taiwanese government decided that THSR would be done as a build-operate-transfer (BOT) franchise: THSRC is in essence a private financial corporation taking credit and contracting out work. It won the franchise in 1997 with the promise of requiring zero public spending.

However, following the 1998 Asian Crisis, THSRC had trouble collecting the needed capital on financial markets. So it took what it could. That included hidden state support in the form of credit and capital from state companies, a 10% share sold to the Japanese construction companies – and credits from banks, including credits with above-market variable interest rates.

A decade later, when the global financial crisis hit, THSRC (which just began to carry passengers) was like American home-owners who took out subprime mortgages: its annual interest payments exploded. This outcome should have been expected with a certainty: exposed to the boom-and-bust, bubble-and-burst financial markets, a mega-project with such long-term financing is bound to face bad times, too. And even had times remained good, the financing was more expensive than government borrowing.

Even before interests, THSRC also faced high depreciation costs, due to an unreasonably short depreciation period of 26.5 years (while some THSR fixed assets might last 80–100 years). Furthermore, the unrealistic initial ridership forecasts can be seen as the result of wishful thinking by players eyeing quick success in private economy terms of quarterly results.

Though interest rates fell again in 2009, passenger numbers stagnated due to the downturn, thus overall losses were only narrowed, and the problem remained. But shareholders were neither willing to increase their own investments, nor to involve new foreign capital to refinance. The Taiwanese government had to step in. However, rather than admit that the BOT model was a bad idea and nationalise the company outright, it took over THSRC indirectly: it took majority control (directly and via state-owned companies) in November 2009.

To solve THSRC's twin financial problems, the government underwrote a massive refinancing loan from state-dominated banks, with lower interest rates and longer maturity; and allowed the company to apply depreciation charges that are a function of traffic volume.

In a further parallel to the Global Financial Crisis, on the sidelines of the government intervention, media and politicians engaged in a more symbolic name-and-shame campaign against "fat cats" receiving high salaries from THSRC, leading some managers to accept 50% pay cuts.

Balance sheets

I made an effort to represent the evolution of THSRC's finances graphically. Let's start with the first year, 2007, when net loss dwarfed ticked income:

Earnings before interest, taxes, depreciation and amortisation (EBITDA) is a metric frowned upon in recent times, however, in our case, it is useful to demonstrate that just running high-speed trains is a highly profitable business, even at the initial low traffic levels. The costs are dominated by fixed costs (what was spent on building the rail line, stations, and trains), hence the negative operating income ( = earnings before interest and taxes: EBIT). You also see that interests are of the same magnitude as depreciation!

A year later, as ridership climbed, ticket revenue increased – but, with the deepening of the global financial crisis, so did interest rates. Thus, operating loss narrowed, but net loss (which includes financial revenues and expenses) was almost as bad as the year before:

Recession hit Taiwan from the second half of 2008, with effect on traffic through 2009. But THSRC managed to hold income steady (more later), and then came the first improvements regarding depreciation charges and interest rates (see previous section). As a result, in 2009, the operating income turned positive and the net loss was cut radically:

With the refinancing deal and further traffic increase, this year shall show further improvement.

Turnaround?

One thing THSRC doesn't have to fear is airline competition. On all relations it serves combined, THSRC achieved an air/rail market share of 89% in the first year already, increasing to over 99.9%(!) in 2009: parallel domestic flights were de facto eliminated, with a Taipei–Kaohsiung alibi service remaining.

The effect on road traffic in available statistics for toll roads (most of which are the two highways paralleling the HSR line) is less clear: if it is there, it is overlaid with the effect of economic growth and recession; and the stats give only an aggregate figure for relevant buses and irrelevant (non-semi-trailer) trucks.

In THSRC's own traffic numbers, the one trace of the recession is the interruption to steady year-on-year growth in the middle of 2009.

Some irregular effects also indicated:

• larger grey-framed squares: the Chinese New Year holiday, which shifts between January and February
• smaller blue-framed squares: safety shut-downs due to major typhoons and earthquakes

As for this year: you see that growth picked up again. With that, revenue from tickets. Though Taiwan predicts an economic slowdown in the second half of 2010, at year end, I expect THSRC very close to break-even.

Let's hope that once THSRC turns a profit, management can be convinced to cut fares or expand discounts further. At least they now give more thought to commuters with newly introduced non-reserved periodic tickets.

A THSR train left one of the many shorter tunnels south of Hsinchu station, near Toufen, on 2 September 2007. Photo by Ernesto JT from Flickr under CC-by-nc-sa 2.0

:: :: :: :: ::

Check the Train Blogging index page for a (hopefully) complete list of ET diaries and stories related to railways and trains.