Some lessons from an investment in Serco

Serco is the first investment to produce a net loss for the UKVI portfolio. As inevitable as this was, selling at a loss is still an unpleasant thing to do. But rather than cry over spilt milk, my task now is to try to understand why the milk was spilt in the first place in order to avoid spilling it again in future.

The fact that Serco has performed badly is not a complete surprise; when it joined the portfolio in May 2014 it was already in trouble.

In 2013, an independent audit by the UK Ministry of Justice found irregularities in the billing of one of its UK Government contracts. Initial findings suggested the company could have benefited by tens of millions of pounds.

As a result, Serco launched its own review as the Government expanded its investigations to include other contracts. By late 2013 Serco’s CEO was gone and shortly afterwards a Serious Fraud Office investigation began.

My assumption at the time was that Serco stood a decent chance of recovering well. However, that’s not what happened.

One problem followed another until the dividend was suspended, a £550m rights issue was launched and the shares fell by 50%. The only plus is that Serco has provided important lessons that will improve the stock selection process going forward.

Buying: A long-established, fast-growing company with government clients

When Serco came up as a potential investment I was, as usual, after a high-quality company at an attractive price. In terms of being a quality company, Serco had a long track record of consistent growth, having grown over the previous decade by almost 15% a year.

It also had financial obligations that I thought were acceptable, although not insignificant, with a Debt Ratio of 3.8 (5 is the maximum I’ll allow) and a Pension Ratio of 5.8 (where 10 is the maximum I’ll allow).

The chart below shows Serco’s excellent financial performance up to the 2013 annual results.

As you can see, it’s a picture of smooth and steady progress, which is exactly what I like to see. What made Serco really attractive was its unassuming valuation. With a PE10 ratio of just 13.4 and a dividend yield of 3.1%, its multiples were no higher than average, yet Serco’s consistent growth made it appear to be a far better company than average.

Yes, Serco had problems with its UK Government contracts, but as a value investor, I have to be willing to invest in difficult situations because it is usually those situations where the best value is to be found.

In fact, most of the portfolio’s investments that have done exceptionally well have done so because the companies, and their share prices, have rebounded from difficult situations that other investors have avoided; so the fact that a company is having problems is not necessarily a reason to avoid it.

After analysing Serco, I thought it seemed likely that the company would face a tough year or even several. But I also thought there was a good chance its problems would turn out to be molehills rather than mountains.

As a result, I added Serco to the portfolio with a position size of just over 3% and added a similar amount to my personal portfolio.

Holding: No knee-jerk reactions, but no buy-and-hold “forever” either

Unfortunately, Serco did not recover quickly. In fact, it has faired far worse than I expected, so the question now is:

Why did such a seemingly minor issue (some serious but apparently isolated problems with contracts relating to one client) balloon into a major crisis?

To answer that question, I’ll quickly summarise my view of the situation.

Serco’s business is built around large, long, complex service contracts. It operates prisons, immigration centres, railways and other things that require the hiring, training, organisation and management of lots of people. It takes services that are typically provided by national governments and re-engineers the component processes to provide the same service more efficiently and cheaply.

This is a competitive market and there is a great deal of pressure to “win” contracts, which can sometimes lead to “underbidding”, or my preferred term, “suicide bidding”, where a company puts in a bid for a contract which is below the expected cost of providing the service.

As daft as that sounds, it isn’t always a bad idea. Sometimes the contract can be made profitable through performance-related bonuses, such as completing various tasks early or to a higher standard, or through follow-on contract wins with the same client.

The practice of bidding low and underbidding is probably why Serco had been so aggressive in billing the UK Government. Having won the contract, it then needed to make it profitable somehow, and somebody thought aggressive billing was an acceptable way to do it. The UK Government thought otherwise.

Unfortunately, these practices extended into many other contracts. During Serco’s internal review, it became clear that several large contracts would no longer be profitable and that other contracts would become less profitable as the level of service provided would have to be improved.

Another nail in the coffin (although not yet the final one) was the knock-on effect of bad publicity from the company’s over-billing of the UK Government. Some potential and existing clients may have become wary of Serco and as a result, the company won fewer new contracts, renewed fewer contracts, or had to accept even worse terms than before in order to become the winning bidder.

As a consequence of losses on existing contracts, lower margins on new contracts and fewer contract wins and renewals, Serco’s profits collapsed in 2014. That in turn led to the dividend suspension and major rights issue.

So that’s the situation as I see it. The next step is to look for things that were visible beforehand that would have suggested that Serco was a “value trap” rather than a good turnaround candidate.

Value trap signal 1: Weak profitability

Because contracts are won through a process of competitive tendering, the lowest bidder is often the winner. As a result, companies like Serco compete largely on price, which means that profits are very thin.

Back when I looked at Serco in May 2014, I didn’t look at profitability at all, so I completely missed this fact. It was the events at another holding (Tesco) and a subsequent article by Terry Smith in the Financial Times, that prompted me to look at profitability via return on capital employed (ROCE).

After some research on the subject, I added ROCE as a ranking factor to my stock screen in November 2014, using 10-year median post-tax ROCE.

As Serco had a low ROCE of 8.1%, which is only slightly above my minimum acceptable value of 7%, it would not have ranked so highly on the UKVI stock screen if ROCE had been part of the ranking calculation at the time.

A lower price would have been required for the company to look attractive enough to buy, so I wouldn’t have bought it in May 2014. Perhaps the stock wouldn’t have looked cheap enough until late 2014, by which time it may have been more obvious that trouble was coming, but I don’t know for sure.

Value trap signal 2: Highly indebted cyclical company, dependent on large, complex, long contracts

Weak profitability alone wouldn’t have automatically ruled Serco out as an investment though, so I may still have invested at some point (although at a lower price). On the other hand, something that would definitely have ruled it out was its excessive levels of debt.

Debts are something I have always looked at, and Serco was no different. When I reviewed the company it had a Debt Ratio of 3.8, which was less than the maximum of 5 I was applying at the time.

However, the Debt Ratio in early 2014 was an earlier version of the one I use today. It was more lenient and didn’t take into account whether the company was cyclical or defensive in nature (cyclical companies will typically be able to handle less debt than defensive companies because their earnings are usually more volatile).

In late 2014 the old Debt Ratio’s shortcomings became apparent, as Serco, Tesco and Balfour Beatty (all holdings at the time) had significant problems.

My response, as usual, was to attempt to improve the underlying investment strategy rather than sell the shares.

That meant changing the Debt Ratio to be more cautious, especially for cyclical sector companies (like Serco, which operates in the cyclical Support Services sector), and that fix has been in place since in December 2014.

Under the new and improved Debt Ratio, Serco’s debt burden looked far less sensible.

With the old approach, it had a Debt Ratio of 3.8 in 2013, which was fine, but the new Debt Ratio gave the company a score of 5.2, which is above the maximum of 4 which I’ll allow for cyclical companies.

If I had been using the new Debt Ratio when I first reviewed Serco, it would never have made it into the portfolio in the first place, and I think that’s the right outcome.

The main reason Serco had to launch a £550m rights issue was to reduce its debts. Under the old Debt Ratio, its debts appeared to be manageable, but clearly that wasn’t true. The new Debt Ratio more accurately indicates that Serco’s debts were not prudent.

Another improvement to the investment strategy, introduced in January 2015 largely as a consequence of Serco and Balfour’s problems, was the addition of a series of “value trap” questions.

These include a specific question on whether or not a company relies on large contracts. If it does (as Serco does) then other factors, such as debts, pension obligations and so on should be viewed with an even more cautious eye.

Value trap signal 3: Large pension obligations

When I updated the Debt Ratio to be more cautious I also changed the Pension Ratio as well, because they both work in essentially the same way by comparing a financial liability to average earnings.

Under the old approach, Serco’s Pension Ratio was 5.8, but the new more cautious approach gave the company a different result.

In the 2013 annual results, its defined benefit pension obligations were quoted as being £1.4bn. At the same time, the company had earned an average of £167m in post-tax profits during the previous five years. That gave Serco a new Pension Ratio of 8.2, which is much closer to my maximum of 10 than the old value of 5.8.

Although the pension scheme didn’t have a large deficit, the company had made “special payments” into the fund totalling £80m in the previous five years, so the scheme was definitely a drag on performance and represented a significant risk.

Serco may have still passed the new Pension Ratio test, but the fact that the company had both high debts and relatively high pension obligations obviously puts it at more risk than if it had just one or the other.

Because of this, I recently added a new rule of thumb to my company analysis process. The new rule is:

• Only invest in a company if the sum of its Debt and Pension Ratios is less than 10

You can read more about how I came up with that rule here.

If I had used that rule during my analysis of Serco, I would have found that its combined Debt and Pension Ratio was 13.4 (i.e. 5.2 + 8.2), which is simply too high.

That would have been yet another reason to avoid it.

Value trap signal 4: Too many large acquisitions

Yet another reason to be wary of Serco was the amount it had spent on acquisitions in recent years. Looking at a company’s acquisition history is a relatively new addition to my company analysis process, one I added as part of those new “value trap” questions.

One definition I use is that if a company has spent more on acquisitions than it earned in that year, then that is a “large” acquisition. Any large acquisitions should be investigated to assess their potential for disrupting the core business.

If a company has made several large acquisitions or spent more on acquisitions over the last decade than it produced in total profits, then I would seriously consider skipping the company altogether.

Large acquisitions are often hard to integrate, are a distraction from the core business, and are often used to grow earnings per share when the core business either cannot grow or produces weak returns on investment. All of those are features I’d rather avoid.

In Serco’s case, it made “large” acquisitions in 2005, 2008 and 2011. Over the 10 years to 2013, it spent £1.1bn on acquisitions compared to total post-tax normalised profits of £1.2bn.

With hindsight, I would say that was too much spent on acquisitions, especially given that Serco is a cyclical company, dependent on large contracts (both of which make the company even riskier).

Selling: Suspended dividends and a highly uncertain future

At the start of June, Serco was the lowest-ranked stock in the portfolio, using the ranks assigned by my stock screen.

As part of my investment strategy, I will sell, every other month, the lowest-ranked holding (i.e. the stock with the weakest combination of growth, income, value and quality) and replace it the following month with the highest-ranked stock in the FTSE All-Share that I don’t yet own (after a detailed and lengthy analysis, of course).

The only reason I won’t sell the lowest-ranked stock is if I think there is some compelling reason to hold onto it, despite its low rank. In Serco’s case, I don’t think there are any compelling reasons to hold onto it.

As a result, I sold the entire Serco position at the start of this month, from both the UKVI portfolio and my personal portfolio. I’ll be looking to reinvest the proceeds next month.

In summary, Serco may not have been a financially profitable investment, but at least it has been educationally profitable. It has led to improvements in my investment process, which will hopefully lead to improved investment returns in the long run.

You can’t win them all…

Although this investment in Serco has performed poorly, the UKVI portfolio has outperformed the FTSE All-Share over the past year (and three years, and from inception), which to me is evidence that it isn’t a good idea to focus on the results of a single investment.

What matters is that your portfolio does what it is supposed to do, which in the case of my portfolio is to produce a market-beating dividend yield, with higher growth and lower volatility than the FTSE All-Share.

Some holdings will do badly (such as Serco, with its 50% loss in one year) while others will do well (such as Cranswick, which has gained more than 115% in less than 2 years), but ultimately it is the performance of the overall portfolio that counts, not the individual holdings.

“The result of one particular game doesn’t mean a damn thing, and that’s why one of my mantras has always been ‘Decisions, not results.’ Do the right thing enough times and the results will take care of themselves in the long-run.”

Amarillo Slim, Poker Legend (Quote from “The Art of Value Investing”, by Heins & Tilson)

Note: You can read the full pre-purchase review of Serco in the May 2014 issue of UK Value Investor (PDF).