Fundhouse Insights Latest Articles
This article focusses on a ratio known as the ‘cash flow yield’. This a financial ratio that measures how well a company generates cash from its current operations. It is calculated by dividing net cash flow per share by the current market value per share. This ratio effectively measures a company’s cash-generating efficiency.
In theory an investor could use reported earnings divided by the share price instead of cash flows to get to a similar outcome. However, in reality, earnings are often full of accounting adjustments that may not always reflect economic reality, so cash flow is often preferred because it is ‘cleaner’ (and harder to manipulate). Dividend yield is perhaps more precise than cash flow, because it represents an actual physical payment to the investor. But, like earnings, it can be adjusted by accountants and executives to serve a ‘non-economic’ purpose. For example, shareholders like steady dividends and companies may prioritize dividend consistency when they have a more cyclical earnings stream. So, given some of these challenges with earnings and the dividend yield, we wondered how cash flow would hold up as a metric of future returns in its own right.
Cash flow yields are calculated with reference to the share price (like dividend yields). This means that they are often highest when share prices are depressed. This is because the denominator in the cash flow yield equation is the share price. The theory is that because cash flow yields are a proxy for cheapness (with a higher yield indicating a greater degree of future), higher cash flow yields should lead to higher returns, all else being equal. This is because a higher yield means that you are paying less per unit of cash that is flowing through the business.
To assist with the analysis, we introduce two caveats: Firstly, we will be looking at aggregate market cash flow yields, rather than company specific yields. So, in effect, we are testing whether the cash flow yield of the US equity market is predictive rather than the cash flow yield of, say, Amazon. Secondly, time horizon is important. Over what period should we judge the efficacy of cash flow yields as a predictor of returns? Should it be over one, three, five or 10 years? If cash flow yields are a strong indicator of forward returns over 10 years only, is this sufficient? We test each of these time horizons below.
We start by testing the correlations between free cash flow yields and long-term returns over different time horizons. What we find is that cash flow yield is highly correlated with long-term returns. However, correlations decrease in a fairly linear fashion as you shorten time horizons. This is a consistent finding over both the UK and US markets and shown in the table below.
returns are monthly and in local currency
Next, we look at a chart that shows the spot US cash flow yield at a certain point in time, compared against the 10-year lagged return of the US market. The UK chart looks very similar.
Figure 2 demonstrates that a strong relationship exists between the starting US cash flow yield and the lagged return of the US market. The averages of both are also very similar, with the cash flow yield averaging close to 14% and10-year rolling returns averaging11%. This means that, on average, the 10-year annualised returns are equal to 78% of starting cash flow yield. There is only a single period where the returns of the US equity markets exceeded the starting free cash flow yield. This being through the late 80s / early 90s. This is because the subsequent 10-year return included the tech bubble.
The cash flow yield of the US market declined to 6% during the heights of the tech bubble. Given that the cash flow yield of the US market averages around 14%. And, because it shifted so substantially from this mean (by over two standard deviations), we do think it was a strong sell signal. If you had reduced your US equity exposure at this point, then you would have avoided the subsequent 10-year negative return that the market experienced.
If, for example, we found that over some periods, returns were strong when cash flow yields were at low levels, that would be problematic to using cash flow as a signal. And vice versa, what if cash flow yields were high and returns were subsequently low or negative? Both scenarios represent potential ‘false signals’ from the ratio.
Figure 3 can help us determine whether we have any false positives coming out of the data. What we show, for the US Market only (but UK is very similar), are three return ‘brackets’ for subsequent 10-year per annum return numbers: <5%, 5-10%, and 10%+. We then looked at the minimum and maximum starting cash flow yields for each of those scenarios to see how much overlap there was.
What we found was:
- The highest returns come in periods where the highest starting cash flows occurred, on average, and vice versa;
- Within each category, there was very little overlap between the starting yields. This suggests that there are very few false positives, especially at extremes, when valuations are cheap and expensive (i.e. cash flow yields are high or low).
The data suggests that the cash flow yield is also a good predictor of returns, 10 years out. To further support this statement, Figure 4 shows the spot cash flow yield versus the lagged 10-year US market returns that correspond to that spot yield. The lagged returns are quite range-bound; they are almost always below the starting cash flow yield and are also a fairly consistent proportion of the starting yield.
Is there a stable relationship between the starting cash flow yield and subsequent return?
The starting cash flow yield does seem quite predictive of future returns. 70%of all observations showed that the 10-year returns approximated between 60% and 100% of the starting free cash flow yield 10 years before. And, where the relationship broke down, it was exclusively in periods like the late 80s when starting cash flow yields were unsustainably high (30% plus) or during the tech bubble when subsequent returns were negative. And, both periods are when markets were either very cheap or very expensive.
- Starting cash flow yields have better predictive value the longer the time horizon. We found that the starting cash flow yields ended up predicting returns over the next 10 years quite well. 70% of the time, returns generated ended up being between 60% and 100% of the starting free cash flow yield;
- The starting cash flow yields are highly correlated with subsequent 10-year returns from US and UK equities;
- Outlier cash flow yields lead to outlier returns 10 years later in the US market. Outlier cash flow yields appear to be at 6-8% (indicating very expensive markets) and 11-20% (indicating very cheap markets).
Despite these conclusions we would point out that there are exceptions to the rules outlined above. When conducting asset allocation, it is important to build a framework that uses a variety of fundamental and valuation based measures to get a clearer picture of market conditions. At the end of 2013, for example, cash flow yields in the US were at 9.8%. If this was viewed in isolation, then an investor would have significantly underweighted their US exposure. From the end of 2013 to the end of July 2019 the MSCI USA outperformed the MSCI World Ex-USA by 60%. Using a broader set of metrics would have protected an investor from blindly following a single indicator.