When I put the Nifty Thrifty portfolio to bed a year ago, I hoped it would blossom into a beautiful princess. When I woke it for its annual update, my hopes were dashed. To mix my fairytale analogies, it hasn’t become a sleeping beauty. It’s becoming an ugly sister.
Before speculation about what the portfolio is performing so poorly, I must first explain why substantial growth in the last year and since I made the first selections seven years ago is not good enough. I must also explain why I was confident the Nifty Thrifty would do better.
The total value of the Nifty Thrifty portfolio has increased 19 per cent to nearly £55,000, from just over £46,000 a year ago. Including dividends, and having made deductions for broker fees and stamp duty, it’s worth 83 per cent more than the £30,000 of imaginary money I put to work in the stock market in June 2010.
Cause for celebration, you might think – except that the benchmark, an investment in the accumulation units of a FTSE 350 index tracker, has performed better. The tracker increased by 26 per cent to just over £56,000 over the past year. It’s grown 87 per cent in value since June 2010. Although the Nifty Thrifty has never managed to achieve a really convincing lead over its benchmark, it has now fallen marginally behind it.
The tracker is a very appropriate stick to beat the portfolio with. The Nifty Thrifty is a selection of 30 shares from the same group – 350 of the largest firms listed on the main market of the London Stock Exchange. Dividends are included in the performance of the tracker, as they are in the Nifty Thrifty portfolio. The tracker, though, emulates the performance of all 350 shares in the index.
Since selecting shares annually takes much more effort than buying one fund and holding it, that effort should be rewarded by superior performance. One or two consecutive bad years shouldn’t discourage us. No system or investor I know of has beaten the stock market every year for a long period of time, but as the short term turns into the long term, effort should pay off. That’s why it’s so disappointing to be reporting that, after seven years of selecting shares, the Nifty Thrifty is losing in the battle to beat the index.
Blame it on Brexit
It’s tempting to blame Brexit. The chart shows the Nifty Thrifty crashed in value by about £8,000 in just a few days after 23 June 2016, the day of the referendum on membership of the EU. The index tracking fund only declined by about £2,000.
While the Nifty Thrifty recovered in absolute terms, it never fully recovered relative to the index.
Many of the shares experienced double-digit percentage share price declines on the days after the vote, particularly those focused on the UK in sectors likely to do badly out of Brexit. They included estate agency Foxtons, property developer St Modwen and retailer Halfords, all of which have remained below their pre-Brexit highs.Perhaps it’s just bad luck. The Nifty Thrifty was invested in the wrong stocks at the wrong time.
Comforting though these thoughts may be, they miss the point entirely. The Nifty Thrifty selects shares using an algorithm that is meant to uncover good companies at cheap prices. The portfolio may respond badly to events from time to time and it may respond better than expected, but over long periods the logic of the algorithm should prevail.
How the Nifty Thrifty works
The Nifty Thrifty algorithm has picked decidedly average investments over the past seven years, but it has used financial statistics commonly believed to identify good companies at cheap prices – statistics that have been tested by academics, investment banks and reputable investors.
The first is return on capital. This is analogous to an interest rate on a bank account. The amount of money you receive in interest payments doesn’t tell much about how good the bank account is unless you compare it to the amount you saved. Currently, you would do well to get a 1 per cent return from a bank, which is to say you get paid one hundredth of your savings a year in interest.
A company invests in buildings and equipment to earn a return, which it receives when it sells goods and services. The return is the profit the company earns, but just like bank interest payments, the size of the profit isn’t sufficient to tell us whether a business is a good one. It depends on how much the company invested.
By doing the same calculation – comparing the company’s return (profit) on capital (money invested) – we can derive its interest rate. If a company earned a profit of £10,000 in its most recent financial year and it had used buildings and equipment worth £100,000, it earned a 10 per cent return on capital.
It’s probably worth being in business to earn a 10 per cent return on capital, but the companies selected for the Nifty Thrifty earn considerably more than that (see candidates table below).
The higher the return on capital, the better the company is at making profit. Just like good bank accounts, good companies are better at making money.
While profit belongs to shareholders, our return is rarely the same as the company’s return on capital. It depends on the price we pay.
To calculate the shareholders’ return, or interest rate, we compare the market value of the whole business to its annual profit. A company worth £500 million that earned a £50 million profit in the last year has an earnings yield of 10 per cent, which, compared to alternatives like a savings account, looks very attractive. Even so, many of the shares selected for the Nifty Thrifty promise an even higher return.
The lower the price of the investment compared to the return it makes, the higher the earnings yield. Shares with high earnings yields are probably good value, or even cheap.
The Nifty Thrifty algorithm ranks each company by its return on capital and earnings yield and then re-ranks them so the companies with the highest combinations are at the top. With some justification, then, it can claim to be finding good companies at cheap prices.
A third statistic, the F_Score, is a trip switch protecting the system. The F_Score is in fact a suite of nine financial statistics that can sometimes discriminate between weakening companies and strengthening companies. A company must not only be profitable, it must also be more profitable than it was the previous year, for example, if it is to score full marks. The Nifty Thrifty requires selections to have F_Scores of five or more out of nine.
That’s the theory. In practice there are complications
Accounting rules are complex and companies apply them in different ways, which makes comparison using statistics derived from accounting numbers challenging. One loophole of particular concern, given the number of retailers selected by the Nifty Thrifty, is the treatment of operating leases.
Operating leases are rental obligations that tie businesses in for many years. In practice, the obligation means the company owns an asset – a lease on a shop, say – but because it’s not the legal owner of the property, it’s not required to recognise the value of the lease on its balance sheet. The leases do not count as capital invested in the business and so return on capital is lower than it should be.
The rules are changing, but the Nifty Thrifty may have been favouring companies with lots of hidden financial obligations. It doesn’t sound like a recipe for success.
There are more general concerns. I derived the Nifty Thrifty from research by investment bank Morgan Stanley and Joel Greenblatt, a legendary investor and fund manager. Their tests on historical data were conducted in the last decade. Based on a period of rapidly growing corporate profitability, they promised much higher returns than I have achieved. Since the financial crisis corporate profitability has stagnated, and it’s possible the algorithm is less effective.
There is also a third explanation for the algorithm’s failure so far: human error. I’ve tinkered with the Nifty Thrifty algorithm over the years in an attempt to overcome deficiencies in the data. I’ve also changed the data provider, and then undone my tinkering.
Last year, I switched the portfolio from quarterly updates to annual updates, which means some shares remained in the portfolio nine months longer than they should have before they were replaced.
I’d be surprised if these were big factors, but in case I’m the problem, now the algorithm is settled again we’re going to give it another three years, until its 10th birthday, to see what it can do.
In this year’s update the algorithm has reselected 10 shares and rejected 20 to make way for new selections. It continues to buy up retailers and housebuilders, which gives me goosebumps.
I shouldn’t be worried, though. The purpose of investing mechanically is to take away the emotion. I’d find that easier, of course, if the algorithm had been right more often in the past.
Subscribe to Money Observer Magazine
Be the first to receive expert investment news and analysis of shares, funds, regions and strategies we expect to deliver top returns, plus free access to the digital issues on your desktop or via the Money Observer App.Subscribe now