I have previously referenced my support for the idea, advocated by Keynes and Adam Smith, that low long term interest rates are a desirable stance for monetary policy. The claim about the effect of low rates is two fold:
(1) Low rates reduce the cost of investment and so encourage it.
(2) Low rates reduce the yields required to pay back debt incurred, and hence encourage more sustainable, less speculative investments. To phrase it conversely: high rates push people into speculation as they attempt to recoup the money they owe.
Commenter Roman P. is not convinced by this argument. I am willing to admit I have, thus far, provided insufficient evidence for this, mostly due to lack of data. However, I have assembled what data I can below, and believe it offers broad – though not definitive – support for this hypothesis.
A few caveats. First, let me establish clear criteria for what I consider to be ‘low rates.’ John Maynard Keynes wanted the long term interest rate to be as low as 2.5%; he even remarked that 3.5% would be too high for full employment:
There is, surely, overwhelming evidence that even the present reduced rate of 3½ per cent on long-term gilt-edged stocks is far above the equilibrium level – meaning by ‘equilibrium’ the rate which is compatible with the full employment of our resources of men and equipment.
For most of the data, the rate is above even the 5% that Adam Smith thought should be the cap, lest the capital of a country be “wasted.” Obviously we shouldn’t believe something simply because Keynes and Smith did, but hopefully the evidence I present below will lend some credibility to their arguments.
Second, what matters will not be just the interest rate; expectations – and the realised trajectory – of the interest rate will also be important. If the rate is rising then it will have a similar impact on investment decisions as an already high rate. If the Central Bank (CB) is committed to a policy of low rates, then it will be far more stabilizing than if rates happen to hit a low point and subsequently bounce back. We do have a test for an explicit low rate policy: the post-WW2 arrangements. It is common knowledge that the stability in that period was unprecedented.*
Third, let me make obligatory correlation =/= causation remarks. Nevertheless, correlation at least gives us a clue about causation. A further clue is if what we think is the causal variable (interest rate) moves first, and the dependent variable (growth) moves second. It is also true that we have a valid theoretical link for our causation. Lastly, it is empirically verified that businesses consider long term rates the most important interest rate in their borrowing decisions.
So what does the evidence look like? Let’s start by taking a look at the ‘Prime Loan Rate’ in the US for the second half of the 20th century. This is the interest rate banks offer to their most stable customers, mostly big businesses:
Every single recession is preceded by an increase in rates. Not every rise in interest rates create a recession – there is one peak without a recession from around 1983-4. However, this may well be explained by movements in the base rate; it dropped from 11% to 8% in that period. By the next recession it had settled at about 6%; that recession seems to have ended when it was reduced down to 3%.
The data for the prime rate only go as far back as 1955, so I’ll use two of Moody’s corporate bond measures for the first half of the 20th century:
Again we observe a similar pattern with rate increases and recessions. Furthermore, the high rate, high volatility period between WW1 and WW2 sits in stark contrast with the low rate, low volatility period post-WW2. It’s interesting to note that rates – though high, relative to our benchmark of 2.5% – were not that high during the stock market boom of the 1920s. Certainly the spike in rates after the first crash is what seemed to bury the economy.
Update: commenter Magpie helpfully pointed out that the Moody’s data could be lagged, which is why it falls inside recessions instead of before them. Indeed, this is what we see when we compare it to the prime rate post-WW2: the spikes are late.
Overall, it seems high or rising rates accompany periods of substantial periods of economic turmoil, else periods where speculation is rampant and bubbles are building up. It is possible the speculation fuels further rises in the interest rate as the perpetrators become overconfident about their potential gains – a positive feedback loop.
Clearly the central bank does not control corporate borrowing rates directly. However, it does control government bond rates, and I would argue that this rate, as a benchmark, has a significant impact on other interest rates in the economy. Indeed this is borne out by the data:
(For a more comprehensive, but uglier, graph of the correlation between government and corporate bond yields, see here).
A central bank committed to low rates could help quell this, as we observe in the data post-WW2. Naturally, such a policy requires a degree of monetary autonomy that central banks have not had since the Bretton Woods system was in place, else rates be disrupted by international flows.
I think the evidence presented here is a blow to the ‘too low for too long‘ meme that pervades discussion of the crisis. There seems to be a belief that low rates are somehow ‘artificial’ (relative to what, exactly?) and we need to ‘get back to reality.’ In fact, it seems that ‘checking’ a bubble may both fuel speculation and needlessly invalidate potential investments, hence creating the situation that the central bank purportedly wanted to prevent.