Michael Pettis: What Does A “Good” Chinese Adjustment Look Like?

the way in which China will adjust. There are two ways to think about the “cost” of financial repression to net savers. The least sophisticated but easiest to explain is simply to look at the real return on loans and deposits. In this case you would subtract the appropriate deflator from the lending or deposit rate.

In the US we usually use CPI inflation as the deflator, but for many reasons this won’t do in China. Consumption is a much lower share of GDP in China than in the US or anywhere else, so that it is less “representative” of economic activity, and there is anyway a great deal of dispute about the rate at which the consumption basket is actually deflating (Chinese households seem to think it seriously understates inflation). I prefer to use the GDP deflator, which until about 3-4 years ago was in the 8-10% region and currently runs around 1-2% or even less, depending on the period you are looking at.

Using the GDP deflator suggests that the real rate for savers has been very negative for most of this century until the past three years – with savers implicitly losing perhaps as much as 5-8% of their real savings every year. It also suggests that the real rate for borrowers has also been negative, perhaps by 1-3% for most of this century. Clearly these interest rates are too low, especially for a very volatile, poor, and rapidly growing economy.

The more appropriate measure of financial repression is not the deflator, whichever one we choose to use, but rather very roughly the gap between the nominal lending rate and the nominal GDP growth rate, the latter of which broadly represents the return on investment within the economy. Until a few years ago nominal GDP grew at around 18-21% while the lending rate was around 7%.

This is a huge gap. In this case the “cost” of financial repression to households was the gap between nominal GDP growth and nominal lending rates, plus an additional 1-1.5% to account for the larger than normal gap between the lending rate and the deposit rate. This is because in China the gap between lending and deposit rates during this century has been much higher than in other developing countries, probably as part of the process of recapitalizing the banks after the last banking crisis at the turn of the century.

If you multiply the sum of these two gaps by the total amount of household and farm deposits (very roughly around 80-100% of GDP a few years ago, when I last checked), you get an estimate of the total transfer from the household sector to banks and borrowers. Because I think China’s nominal GDP growth has been overstated by a substantial amount because of its systematic failure to write down bad loans, I usually have subtracted 2-4 percentage points from the nominal GDP growth rate before I did my very rough calculation. This was how I got my 5-8% of GDP estimate for the amount of the annual transfer from households to savers. This of course is a huge transfer, and can easily explain most of the decline in the household share of GDP over this period.

It is worth noting by the way that a recent widely-discussed study by Harry Wu of the Conference Board claims the China’s average GDP growth from 1978 to the present was not 9.8% but rather 7.2%. The main reason for the revision, according to Wu, is that the GDP deflator had been significantly underestimated which, if even partially true, means real interest rates were even lower (more negative) than I have assumed.

There is some controversy about whether it is true that the nominal lending rate should be broadly equal to the nominal GDP growth rate. In fact most studies of developed countries suggest that over the medium and long term this is indeed the case. UBS tried to show that this was not applicable to China and did a study several years ago showing that among developing countries this relationship didn’t hold. Their studies suggested that among developing countries nominal lending rates had on average been around two-thirds on nominal GDP growth rates (although China, at around one-third, was still well below anyone else’s at the time).

I had a real problem with their sample of countries however. Their sample included a lot of small OPEC countries, who necessarily had high growth and low interest rates when oil prices were high, as well as a lot of Asian countries that followed the Japanese development model and themselves practiced financial repression, which of course made them pretty useless as points of comparison. Neither group of countries, in other words, could help us determine what a “normal” interest rate is compared to nominal GDP.

But regardless of the debate, the point to remember is that when the nominal lending rate is much below the nominal GDP growth rate, two very important things happen. First, it helps eliminate capital allocation discipline. If GDP is growing nominally at 20%, for example, and you can borrow at 7% (which was the case in China for much of this century), you should rationally borrow as much as you can and invest it into anything that moves, no matter how poorly thought out the investment. Imagine a totally ineffective investor, or one whose incentives do not include earning a reasonable return on capital, who manages to earn on his investment only half of nominal GDP. This would be a pretty poor use of capital.

Adjusting the repression gap

But with nominal GDP is growing at 20%, this extremely incapable investor still makes a substantial profit by borrowing at 7% and earning 10%, even though his investment creates no value for the economy. His “profit”, in this case, is simply transferred from the pockets of saving households.

Under these conditions it should be no surprise that borrowers with access to bank credit overuse capital, and use it very inefficiently. They would be irrational if they didn’t, especially if their objective was not profit but rather to maximize employment, revenues, market share, or opportunities for rent capture (as economists politely call it).

The second point to remember is that in a severely financially repressed system the benefits of growth are distributed in ways that are not only unfair but must create imbalances. Because low-risk investments return roughly 20% on average in a country with 20% nominal GDP growth, financial repression means that the benefits of growth are unfairly distributed between savers (who get just the deposit rate, say 3%), banks, who get the spread between the lending and the deposit rate (say 3.5%) and the borrower, who gets everything else (13.5% in this case, assuming he takes little risk – even more if he takes risk).

This “unfair” distribution of returns is the main reason why the household share of income has collapsed from the 1990s until recently. I calculate that for most of this century as much as 5-8% of GDP was transferred from households to borrowers. The IMF calculated a transfer amount equal to 4% of GDP, but said it might be double that number, so we are in the same ballpark. This is a very large number, and explains most of why the growth in household income so sharply lagged the growth in GDP.

This was why financial repression, although useful in the early stages of China’s growth period because it turbocharged investment, ultimately became one of the county’s biggest problems once investment no longer needed turbocharging. For many years nominal GDP growth in China was 18-21% and the official lending rate was around 7%. This, combined with widespread moral hazard, had inevitably to result in both tremendous misuse of capital and a sharp decline in the consumption share of GDP (as the household income share declined) – both of which of course happened to a remarkable degree in China.

In the last year or so, however, the official lending rate has risen to 7.5% and nominal GDP has dropped to 8-9% (and just under 8% in the first quarter of 2014). This changes everything in China. First, it is now much harder for borrowers to justify investment in non-productive projects because they can no longer count on the huge gap between nominal GDP growth and the lending rate to bail them out of bad investments. Of course this also means a dramatic slowdown in economic activity, but because this slowdown is occurring by the elimination of non-productive investment, the slowdown in Chinese growth actually represents higher wealth creation and greater real productivity growth. China is getting richer faster now than it did before, even though it looks like wealth creation is slowing (the difference is in the slower required accumulation of bad debt).

Second, the huge transfer from net savers to net borrowers has collapsed, so that growth in the future must be far more balanced. Over time this means that households will retain a growing share of China’s total production of goods and services (at the expense of the elite, of course, who benefitted from subsidized borrowing costs) and so not only will they not be hurt by a sharp fall in GDP growth, but their consumption will increasingly drive growth and innovation in China.

Interest rates are still too low, but not by nearly as much as in the past, and over the next two years as nominal GDP growth continues to drop, the financial repression “tax” will be effectively eliminated. When this happens solvent Chinese businesses will be forced to use capital much more productively, and slowly they will learn to do so. In that case the PBoC will be able to liberalize interest rates (although not without tremendous political opposition from those that have depended on having great access to very cheap capital for their wealth) without worrying about either the deposit rate of the lending rate surging.

There is, however, one group of wasteful borrowers for whom higher interest rates will not represent a more careful approach to borrowing and investing and these, of course, are borrowers that are already effectively insolvent or otherwise unable to repay loans coming due. In that case as long as they can borrow they will do so, no matter the interest rate. It is not clear to me how many such borrowers exist, but I’d be surprised if there weren’t an awful lot of them. These borrowers can only really be disciplined by constraining credit growth and eliminating government support, including implicit guarantees, but this might not be happening.

One of the ways Beijing seems to be reducing the pain of more expensive borrowing (relative to nominal GDP growth) is to loosen credit in a targeted way. We have heard talk of targeted bond purchases although it is not yet clear what exactly Beijing plans to do.

An article in Caixin suggests that regulators may also be trying to relax the loan to deposit constraint:

The China Banking Regulatory Commission (CBRC) will change the rules for figuring the loan-to-deposit ratio so banks can have more money to lend to businesses, an official with the regulator says. The requirement that the ratio not exceed 75 percent – meaning banks cannot lend more than 75 percent of their deposits out – would remain the same, but the way ratio is calculated would be adjusted, Wang Zhaoxing, deputy chairman of the CBRC, said June 6.

The regulator would consider broadening the scope of deposits to include “relatively stable” funds that are not now used to calculate the ratio, he said. He did not say what the funds could be, and added that a precondition for doing this was keeping the money market stable.

I suspect that over the next few months we are going to get very inconsistent signals about credit control. But as long as the PBoC can continue to withstand pressure to lower interest rates – and it seems that the traditional poor relations between the PBoC and the CBRC have gotten worse in recent months, perhaps in part because the PBoC seems more determined to reduce financial risk and more willing to accept lower growth as the cost – China will move towards a system that uses capital much more efficiently and productively, and much of the tremendous waste that now occurs will gradually disappear. Just as importantly, lower growth will not create social disturbance because Chinese households, especially the poor and middle classes, will keep a larger share of that growth.

So what does “good” rebalancing look like?

It seems pretty clear to me that the great distortions in the Chinese economy that led both to rapid but unhealthy growth and to the consumption imbalance (by forcing down the household income share of GDP) are gradually being squeezed out of the system. One distortion has been the excessively low exchange rate, but after seven years of 30-40% net appreciation against the dollar, the RMB is far less undervalued today than it has been in the past. I still do not agree, however, with analysts who say the currency is actually overvalued and call for a depreciation, nor, more importantly, does the PBoC seem to agree.

Another one of the great distortions that led to China’s current imbalances was the very low growth in wages relative to productivity. This too has improved. The surge in wages in 2010-11, and their continued relatively rapid pace of growth, has reduced this distortion significantly, especially as it is becoming increasingly clear that productivity growth has been overstated in recent years.

Most importantly, with nominal GDP growth rates having dropped from 20% to 8-9%, the greatest of all the distortions, the interest rate distortion, has been the one most dramatically to adjust in the past three years. This is why even though many people I respect are still insisting that China has not really rebalanced, I am moderately optimistic that in fact China is adjusting as quickly as could be expected. Credit growth remains a serious problem, but the forces that put China in the position of relying on excess credit growth have genuinely abated.

And it is this abatement of the great distortions that have caused growth to slow so rapidly, and although we haven’t seen much evidence of significant rebalancing yet, it should take a few years for the effects fully to be worked out. Chinese growth is less dependent than ever on the hidden transfers from the household sector, and these