Tight money is back. Given that $300 trillion in global debt, public and private, needs to be rolled over at higher rates for the foreseeable future, it is important to understand whether, and if so how, debt burdens affect the transmission of monetary tightening to real activity. Tight money also has important implications for returns on equities, bonds and housing. But we’ll leave that for a future dispatch. In what follows, we’ll obtain a quantitative estimate of the effect of higher rates on real activity. Do higher rates, in the context of higher debt burdens, imply a somewhat exogenous negative growth shock for the core of the world economy?
We try to answer this question by interrogating international panel data for advanced economies. We’ll work throughout with the Jordà-Schularick-Taylor Macrohistory Database. We have data on 18 advanced economies for 50 years. We construct (total) debt-to-gdp ratios by adding up the nominal debt of households, nonfinancial firms and states and dividing the sum by nominal GDP. The reason we pay attention to total debt ratios instead of public debt, is that not just public debt but private debt as well has to be potentially refinanced at higher rates. For simplicity, we’ll call the total debt-to-gdp ratio, the debt ratio.
The debt ratio across the rich world has never been higher. By 2017, when our data ends, the median advanced economy’s debt ratio stood at nearly 200 percent of gdp. US debt ratio stood at 168 percent of gdp in the same year. Since then they’ve grown by a fifth, if indeed not a third.
But we’re not interested in levels. We detrend long-term and short-term rates and debt ratios by first differencing. We obtain detrended real gdp and per capita gdp (pcgdp) by dividing nominal gdp and pcgdp by the consumer price inflation index and then detrend by computing annual growth rates. We robustly standardize all variables to have zero mean and unit variance — so the estimated elasticities will be in standard deviation units. We estimate panel regressions using PanelOLS. We always include country fixed-effects and cluster errors by country. We also include an intercept term, as well as a trend term. For goodness-of-fit we look at within R-squared. The response or dependant variable will be growth, as measured by detrended rgdp or rpcgdp. As features or predictors, we include the short-term interest rate, the long-term interest rate, the debt ratio, and an interaction of the debt ratio with the short-term rate. We lag all features by a year. Each of them is of some research interest.
The first regression is of real growth. The within R-squared of this model is 22 percent, which is not bad for a panel spanning 18 countries and a half century. We find that a (standard deviation) unit higher debt ratio shaves off 0.15 units from real growth, a unit higher short rate shaves off 0.10 units, and a higher long term spread shaves off an additional 0.11 units. For reference, 0.10 units of real growth for the median AE is 0.46% per annum and 0.15 units is 0.69% per annum; and a unit hike is 133 basis points (bp) for short rates and 84bp for long rates. We can reject the hypothesis that higher debt ratios make tight money more painful: the interaction term is not significant (p = 0.23). The trend is negative, as expected.
It may be that real growth is confounded by population growth. So, we use rpcgdp as the response instead. Our model can explain 25 percent of the variation within countries. The interaction is again insignificant, suggesting that tight money does not mean lower growth conditional on high debt ratios. The elasticity per capita income to debt ratios is -0.17 in standard deviation units or -0.76% per annum; that of long rates is -0.15 units or -0.69% per annum; of short rates -0.11 units or half a percentage point per year shaved off per capita income growth.
To sum up, tight money means lower output and income growth. We estimate that a half-a-dozen hikes by the hard currency issuing central banks — which now seems like the baseline scenario — will shave off at least half a percentage point off AE growth rates (or 69bp from per capita real gdp growth). Interestingly, we do find a significant negative effect for market rates at the long end (or slope effect) over and above that due to the short rates (or level effect). We find that this duration effect of similar magnitude as the direct effect due to innovations in the monetary policy rate.
Output growth is likely to be even lower because of the recent growth in debt ratios. We estimate that a standard deviation shock to the debt ratio reduces a generic advanced industrial economy’s growth rate by 75bp (or 86bp for per capita real gdp growth). But we do not find any evidence that higher debt ratios interact with higher rates to lower growth rates over and above the lower rates already expected on account of higher aggregate (public and private) debt ratios and interest rates.
The bad news is that the tightening cycle will materially temper economic growth; as will the higher debt burdens across the advanced world. The good news is that it won’t take much bitter medicine to cool the world economy, baseless projections to the contrary notwithstanding. So, there’s no reason to believe The Economist’s scaremongering: “There are few examples of central banks taming inflation without the economy suffering a recession. The last time America’s inflation fell from over 5% without a downturn was over 70 years ago.” We may instead be looking at a shallow hiking cycle with a soft landing. Or so one can still hope.
Why do they need to tighten. The inflation is 3 things gas, trucks and meat. Tightening should come with tax raises on wealth.