This is a response to an article by Montier and Pilkington (https://www.advisorperspectives.com/commentaries/2017/03/27/the-deep-causes-of-secular-stagnation-and-the-rise-of-populism). They identify four main changes in the modern economy that they say caused the current stagnation. The article is typical of the type advanced often nowadays by those who don’t want to accept the superiority of free markets, but realise that real old-school socialism has been thoroughly discredited. They then argue that the so-called “neo-liberal system” has failed, and propose an interventionist/developmental state model. My response hopefully arms the reader to deal with many of these fallacious contentions, and so should have more general application.
The first alleged mistake in economic management is the abandonment of the policy of “full employment” followed after WW I. In support of this the authors suggest that Keynesian policies of government spending ensured full employment during the “golden era” of low unemployment (1948-1970).
This is just nonsense. To start with, here is the graph of all federal spending during the last century, showing how spending rose after the War to 1980 (from about 15% to 24% of GDP):
And here is the unemployment rate over, inter alia, the same period, showing an increase from about 2% to 10%:
After WW I and before the depression (1920-29), spending was relatively low, and unemployment was low too. In fact, unemployment during that period was on average about the same as in the “golden era” of 1948-1970, without the aid of sky-high state spending. In the period after the war the trend line of federal spending was gradually upward (1948-1980) – and so was the curve of the unemployment rate. Reagan reversed the spending trend, and also the unemployment curve trended down (1980-2000).
A simple glance at the curves of federal spending and unemployment in the period 1970 to date will show that as spending increases, so does unemployment. This does not necessarily mean unemployment was caused by spending growth, but it sure as hell does not prove the opposite, namely that spending kept unemployment levels down.
Notably in the modern era, the Reagan years (81-89) were marked by a simultaneous decline in spending and unemployment. The same applies to the period of the Clinton administration – 1993-2001. On the other hand under Bush II, spending grew as did unemployment. Obama reversed the trend again.
The same applies during the most famous stimulus programme, the New Deal, which carried on for almost a decade after 1929, without succeeding in bringing unemployment down below 20%:
Significantly, unemployment fell by 10%, after much of the New Deal had been declared unconstitutional in 1935. The War’s low unemployment was abnormal. Bear in mind the so-called war economy was a case of government deficit spending to manufacture arms and ammunition, while Americans went hungry. It was not productive employment. Equally importantly, after WW II the government slashed spending by 70% (see the spending graph above), immediately restoring normal unemployment patterns.
The authors further suggest that the inflation of the seventies was not caused by excessive state (Keynesian) spending. The above federal spending graph certainly would suggest that spending was what caused it. But the authors say inflation in the seventies was caused by the OPEC oil crisis (and labour troubles. We’ll return to that below). But there are ample studies showing that it is not the case that the oil shock caused the inflation peak of the seventies. But just a look at the inflation rate graph of the period before and after the oil shock (which occurred in October 1973) makes it clear that the
inflation trend began long before, in 1960. Furthermore, the inflation rate was in tandem with federal spending. So for example it peaked at 15% during the War (1945) when spending was at a historical high, and then declined in step with spending declining until the early sixties, when it reached about 2%. At that stage government spending was a relatively modest 18%. Then, from about 1962, both spending and inflation increased, until inflation reached about 11% in the early 1970s. Then it reached another peak of about 14% in 1980, which coincided with a further spending peak (just before Reagan took over) of about 24%.
The argument that inflation was caused by the oil crisis (and not by state spending) would have had more traction if state spending did not also reach unprecedented heights during the seventies. How the authors could say increased spending was not accompanied by rising inflation, and was necessary to maintain low employment, is hard to fathom – especially seeing that over the next two decades the trend reversed, and we see the following declines:
– Spending declined from 24% to 19% of GDP;
– Inflation declined from 14% to 3%;
– Unemployment declined from 10% to 4%.
What the authors do not mention either, is that in 1971 the US abandoned the gold standard, (to which the dollar had been pegged), causing the value of the dollar to drop. That caused inflation as dollars flooded the market. Because the OPEC oil suppliers were paid in dollars (which were losing value fast), they increased the price of oil. So it was inflation that caused the rise in the oil price, not the other way around. What is clear is that the delinking of the dollar from gold added to the inflation woes of the US.
This bit of history makes it clear that there is a link between state spending, inflation and unemployment. At present the unemployment rate is 4.7%, better than the “Golden era” average of the sixties of 4.8% (while state spending in 2016 was 34 % of GDP – down from 43% in 2009 – and inflation at 1.26%. The statement that unemployment “stayed permanently elevated” and needs inflationary, Keynesian spending to reduce it, is just not true.
Decadal Estimates of the Average Unemployment Rate
In the Obama era, federal spending decreased as percentage of GDP, as did inflation.
Not surprisingly, during his tenure unemployment also declined:
The authors’ “solution” to low incomes and high unemployment is NAIBER, which is a scheme whereby the government must employ all unemployed workers at a basic wage, end use their services to perform charity, public works or whatever.
But wait, has that not been tried? O yes, they called it the New Deal. Government in the New Deal era ran up unprecedented debts and barely made a dent in the unemployment rate. Why would this plan be any different? On 9 May 1939, Secretary of the Treasury Henry Morgenthau appeared in Washington before the US House of Representatives Committee on Ways and Means and exclaimed:
We have tried spending money. We are spending more than we have ever spent before and it does not work. I say after eight years of this Administration we have just as much unemployment as when we started … And an enormous debt to boot!
The incentives of the scheme are all wrong. Workers will not be required to be productive, because their employment will be guaranteed. Even workers in the private sector will know that their jobs are safe in the sense that if they were to be fired for not doing their work, the safety net of the state would catch them. The state as employer, for its part, will not need to use their service productively, because, well, it is the state. Any waste can after all be funded by the taxpayer. The state will run up costs and debts. The state’s share in the economy will grow. Wealth creation will suffer, as always happens when the state’s share in the economy grows.
Implicit in the NAIBER solution is that it will be inflationary, because the authors’ criticism is precisely that the modern “neoliberal” system wrongly targets inflation. It follows that what they really want is inflationary state spending. As we have seen above, it leads to serious economic problems. The period of 1960-80 was one where state spending, inflation and unemployment were ramped up in tandem.
Next, the authors blame globalisation, notably immigration increases and international trade, for the stagnation of the economies of the US and other developed countries.
Dealing with immigration first, while it is true that immigration to the US is at a historical high, immigrants’ median income is on the whole lower than US workers’ . At the same time, it rises more quickly than that of US workers, and currently unemployment among immigrants is 3.7% , significantly lower than that of citizens. That indicates that immigrants assimilate well into the labour market, which is a key component of social integration. These figures also suggest that initially immigrants do jobs at salaries that citizens are not prepared to accept, and thus are unlikely to disrupt citizens’ employment opportunities.
What is true of course, is that over the last decade or so, American workers in the lower quintiles have seen their incomes decline as a relative share, and in real terms. The reason is that the American market was skewed by successive interventions by the Federal Reserve since the time of Greenspan, in the form of money printing. The effective negative interest rates maintained by Greenspan, together with anti-discrimination policies in the bond market and the implicit guarantee that banks would not fail, led to the 2008 housing crash. That destroyed incomes of poor and middle class people as unemployment soared and home buyers defaulted on mortgage payments. The next 6 years was followed by a period of Zero Interest Rate Policies (ZIRPS) and Quantitative Easing (QE), fancy names for more money being printed and pumped into shares and bonds. That caused an unprecedented boom in the stock market, benefitting rich people, and resulting in dwindling pay cheques in the hands of the middle class and poor of America . The inflationary policies of Greenspan caused headline inflation to rise from 1% to 5.5% between 2001 and 2008. In the Obama years the Fed did not pump the excess money supply into the consumer market, but into assets. The result was that there was little impact on consumer prices, but the stock market was artificially inflated .
Here is the relationship between the value of assets held by the Fed (which is the result of QE and ZIRP – money printing for short) and the value of the share index:
That was a boon for the rich and the worst blow to the poor. It also devastated productivity, as investment in paper assets do not improve technological investment or training of workers.
That is the real reason for the declining incomes of the poor and middle classes. If the economy had been healthy, investment would have been made in training, machines, information technology and productive assets, instead of share portfolios.
US productivity has surely also been depressed by the poor school system, which, despite the government being one of the most prolific spenders on education as percentage of GDP in the world, has failed to show any significant improvement in reading, maths and science teaching over the past 50 years.
Immigrants have always been part of the American labour market. The rising employment level of citizens (at higher wages than immigrants) suggests that immigrants are not taking their jobs or depressing their income. There is no evidence that immigrants take US citizens’ jobs. Quite the contrary.
There is a reason why manufacturing has declined in developed countries. It has to do with the relative skills level of countries, and wages pay for different levels of productivity. The modern economy is knowledge-based. Countries with high-tech products and services such as information technology, consulting services, research and development, have a competitive advantage. Countries with lower skills levels then have to compete with the advanced economies by performing manufacturing jobs at lower wages. Why do the authors want Americans to work in factories? They should be grateful that the economy has graduated to a knowledge- and hi-tech-based industrial complex. Insofar as a section of the workforce has been left behind, then improve the education system. The US school system is positively mediocre compared to the good systems of the world, such as most Nordic countries, most oriental societies and Switzerland. That is a story on its own, of course. But in short, the US education system fails because it is a centralised, government-driven system in a highly diverse society. All the successful systems in the world are either culturally homogeneous (Shanghai, the Nordic countries, Germany) or highly decentralised or privatised (the Netherlands, Hong Kong, Chile, Finland, Switzerland), or both homogeneous and decentralised (Hong Kong, Finland, the Netherlands, Shanghai).
As regards trade due to globalisation: What would the authors have us do: should countries like the US have less trade with the rest of the world? Presumably not. Then one supposes the problem is the dreaded trade balance, namely that countries like the US import more than they export. The answer to that is simple: If the trade balance is a problem, then increase value for money by improving productivity. Improve education, reward real bricks-and-mortar investment, not paper pyramids on Wall Street.
The authors’ prescription to deal with the effects of globalisation, with a view to repatriating “good” manufacturing jobs back to the developed world, is import substitution by way of targeted subsidies for products that can be manufactured more cheaply abroad – in other words, good old-fashioned protectionism through subsidies. In principle import substitution subsidies are the same as protective tariffs. In both cases the state uses taxpayers’ money to increase the relative price paid by consumers for imports. In the subsidy scheme the taxpayer firstly funds the subsidy, and secondly pays for the inevitably reduced value for money brought about by restricting the competition of imports. In the tariff scheme the consumer pays a tax on imported goods (which adds no value to the purchase) or buys inferior value for money locally. Both are destructive of growth, unproductive and wasteful.
For example, here are 5 countries with high ratings for import tariffs according to the Economic Freedom of the World Index (ie with few or no import tariffs), compared to 5 countries with lower ratings. Both groups come from the first quartile of economic freedom, and save for separating them into low- and high-tariff groups, have been selected purely so as to have about the same degree of economic freedom on average. Any difference in average performance is therefore not due to general economic regulation, the significant difference between the two being the degree of import protection only.
The free-trade group (with little or no import tariffs) grew three times as fast as the other group. This accords with more general evidence showing free trade is overwhelmingly associated with per capita income growth:
It follows that the converse, increased protection, will lead to less growth, which will scarcely be likely to address the problem of lower income for workers that concerns the authors. If anything, there should be more international trade if we are serious about addressing income of poor and middle class people.
The next aspect that the authors tackle is labour regulation, contending that labour flexibility has led to lower incomes in the developing world.
As a general proposition, this is just not so.
Real-world examples making this point are not difficult to find. The following are a few examples of countries with lightly regulated labour systems, taken from the periods shown:
In all these cases the countries maintained low levels of labour regulation, low unemployment and growing incomes.
Historically, a prime example of a country that achieved a spectacular ‘race to the top’ is Taiwan (once again with a very lightly regulated labour market), which performed as follows in the period 19611994:
Countries where labour deregulation took place also support this conclusion. In 1991, New Zealand introduced the Employment Contracts Act (ECA), which significantly reduced artificial statutory support for trade unions, and abolished centralised bargaining structures that prevented employers and employees from concluding their own, individual contracts of employment. Here are the results over the period from inception of the ECA to the implementation in 2004 of the Employment Relations Act of 2000, which reversed some of the deregulation of the ECA somewhat:
• Between 1991 and 2004, unemployment fell from about 11 per cent to under 4 per cent.
• During the same period, the labour-force participation rate increased from 63.8 to 66.9 per cent.
• Most significantly, GDP per capita in real terms increased from $12 230 to $25 420.
Also Denmark makes the point. In the period 1994-96 Denmark deregulated important aspects of its labour market, including hiring and firing rules. The result was that unemployment declined from about 10 percent to about 3 percent between 1995 and 2009, while the country’s per capita GDP grew by 16.5% over the same period .
It is simply not true that deregulation is likely to lead, or has led, to a plummeting of labour standards. Quite the contrary.
As for the notion that union membership increases employment, this is simply false. Study after study has shown that the higher the union density of an industry, region or economy is, the lower the employment growth. See the graph below, showing the union density in South Africa and employment growth/decline 1980 to 2004:
As the union density rose from 1980 onwards, employment growth declined from 5.5% p/a to a negative of -4% in 1988. The two curves run in parallel like rail tracks.
In 1947, the United States passed the TaftHartley Act, which affirms the right of states to enact right-to-work laws. Under these laws, employees in unionised workplaces may not be compelled to join a union or pay for any part of the cost of union representation. In anticipation of Michigan’s adoption of right-to-work laws in 2013, the Mackinac Center for Public Policy in Michigan examined the economic effects of right-to-work status. Unsurprisingly, right-to-work states have far lower union density than non-right-to-work states (8.83 per cent as opposed to 15.59 per cent).
Here are some of the study’s most striking findings:
• ‘According to the Bureau of Economic Analysis, right-to-work states showed a 42.6 percent gain in total employment from 1990 to 2011, while non-right-to-work states showed gains of only 18.8 percent.’
• ‘According to the U.S. Census Bureau, population increased in right-to-work states by 39.8 percent and only 16.7 percent in non-right-to-work states from 1990 to 2011.’
• ‘According to the U.S. Census Bureau, 4.9 million people moved from non-right-to-work states to right-to-work states from 2000 to 2009.’
• ‘According to the Bureau of Economic Analysis, nominal personal income grew by 209.3 percent in right-to-work states and by 148.5 percent in non-right-to-work states from 1990 to 2011.’
For present purposes, the most significant impact is on employment growth. Right-to-work states in the period 19702011 grew employment at a rate exceeding that of non-right-to-work states by 0.8 percentage points per year. Importantly, despite growing jobs roughly three times faster, right-to-work states did not sacrifice in terms of personal income. On the contrary, according to the study ‘states with right-to-work laws enjoyed an annual average increase in real personal income of 0.8 percentage points … compared to what they would have experienced without such laws’.
Montier and Pilkington compare countries internationally by way of a scatter gram to show that countries with high union density have low unemployment. They also show that during the period of low unemployment (1948-1970) union density was higher. This is a classic case of comparing apples and pears. In order to know if unionisation increases employment, we must surely control for relevant factors.
To show how easy it is to make such a generalised point, let’s look at the UK instead, that leads to exactly the opposite conclusion. Here are the graphs showing the historical trends of trade union density, unemployment and inflation in the UK:
Union density, inflation and unemployment reached a simultaneous peak in the early 1980s. That was when Margaret Thatcher stepped in and broke the back of the union movement. As trade union density declined, so did unemployment and inflation.
So yes, poor labour relations contributed to inflation. That in turn was caused by high trade union density. Not only that. As in the US, UK government spending also peaked around 1980:
It is far better, if we want to compare the effect of unionisation, to isolate it as a factor. The easiest way to do that is to compare union plants in one industry with non-union plants in the same industry, as the heritage Foundation did in respect of manufacturing in the US:
And in respect of construction:
So, trade unions are a problem for unemployment. The same applies to labour laws generally. In the United States, the federal government regulates many aspects of labour relations. But over and above the federal laws, each state also has the right to enact local laws that regulate employment. Any state could, for example, enact a minimum wage for the state that exceeds the national minimum wage. This makes America an ideal testing ground for the effect of labour laws on employment creation, because many of the economic conditions in states are determined by federal legislation and circumstances that are shared by all states; and citizens of states are free to move from state to state to find work.
Based on a survey of employment policies conducted in 2009, the US Chamber of Commerce developed an Employment Regulation Index (ERI) to measure the impact of state labour and employment regulation in each of the 50 US states on a scale of 1 to 100, with a score of 100 denoting conceivably the most heavily regulated state. The ERI is based on rankings of 34 measures of state labour and employment policies covering six categories: the employment relationship and the costs of separation; minimum wage and living-wage laws; unemployment insurance and workers’ compensation; wage and hour policies; collective-bargaining issues; and the litigation/enforcement climate. Using the ERI, each state is assigned an overall rank – ‘Good’, ‘Fair’ or ‘Poor’ – as an indicator of the extent to which the state’s labour and employment policies regulate labour. It was assumed that states with a ‘Good’ rating have strong pro-employment policies, while those with a ‘Poor’ rating have policies that inhibit job creation.
In a 2011 report of their findings, the Chamber of Commerce measured the economic impact of regulation on the performance of each state:
‘Applying standard statistical techniques, we found that, when the ERI is inserted as an explanatory variable, our models demonstrate that higher levels of regulation (i.e., higher ERI scores) result in both higher unemployment and lower rates of new business formation, and that the effect is statistically significant at standard confidence levels. Moreover, the magnitude of the estimated effects is substantial. We estimate that, if each state were to achieve a ‘perfect’ score on the ERI, the effect would be equivalent to a one-time boost of approximately 746,000 net new jobs.’
When it came to assigning overall rank, 15 states were ranked ‘Good’, 20 ‘Fair’ and 15 ‘Poor’. In the five-year period from 2009 to 2014, I calculated the average employment growth in each group, which was as follows:
Good 2.72 per cent
Fair 0.6 per cent
Poor 0.49 per cent
Quite clearly it pays to be ranked ‘Good’. Average employment growth in this group is about 4.5 times higher than in the ‘Fair’ group, and 5.5 times higher than in the ‘Poor’ group.
A more dramatic illustration of the value of low regulation is to compare the percentage of jobs created over the same five-year period for all the states in each group combined. The ‘Good’ group’s employment growth was 4.4 per cent, while that of the ‘Poor’ group was a measly 0.6 per cent. In other words, the ‘Good’ states proportionally created more than seven times as many jobs as the ‘Poor’ states did. In actual numbers, the ‘Good’ states created 2 124 734 jobs (starting from about 47 million), while the ‘Poor’ states created only 390 551 jobs (starting from a much larger 62 million).
And in case you were wondering, per capita income growth is also better in the ‘Good’ states:
Good 1.82 per cent
Fair 1.64 per cent
Poor 1.68 per cent
So, when labour law is isolated as a factor, it is and remains a job-killer. That does not mean its effect cannot be ameliorated, as for example in the Nordic countries that have national corporatist negotiations. The difference is, they are all homogeneous societies. No multi-ethnic society like America has ever succeeded with a corporatist model.
The authors further argue that the theory that minimum wages do not suppress employment, is not correct. This argument they support with evidence that the minimum wage as percentage of average wage was higher during the so-called “golden era” when unemployment was low, than when it unemployment was high in the period from 1970 onwards. This is a poor argument. Any economist worth his salt will tell you that isolating one factor – the minimum wage – and arguing that it caused or did not cause unemployment, without controlling for other relevant factors such as the general regulatory regime of the economy, the share of government spending, the education of the society and external shocks, is just a nonsense. For one factor, just look at the share of government spending, the records of which appear above. In the so-called “golden era”, spending was lower than in the high-unemployment era of 1970-80. Then spending declined over the period 1980-2000, during which unemployment likewise came down.
In the real world, minimum wages have proved to be much of a muchness. Everyone agrees that if we imposed a minimum wage in the US of $50 per hour, or in SA of R15 000 per month, the economy would take a severe knock and millions would become unemployed. Yet in the real world minimum wages are set at far more modest levels, resulting in far less drastic effects, or sometimes none at all, and thus ongoing bickering among economists as to their importance. The reason for this is that in most real-life situations it is easy to demonstrate that the impact of minimum wages is at worst, modest. The one reason is that only a tiny minority of workers are affected by minimum wages, namely those marginal workers at the bottom rung of productivity. No country’s economy has come to a standstill because of it. Yet studies consistently show that teenage employment is significantly harmed by minimum wages.
The next graph makes the same point, over a longer term. Note especially the decline in the minimum wage, and the concomitant drop in teenage employment from 1980 to 2000.
Again, one remains mindful of the fact that this is not proof that minimum wages cause unemployment. But it certainly shows that the converse is not true, namely that unemployment is lower when minimum wages are higher.
As for the ratio between CEO and worker pay, and other forms of income and wealth inequality, we have seen the impact of asset inflation at the behest of the Fed (above) and the effect it had on income disparity. The authors are correct to assert that CEOs are typically compensated on the basis of their companies’ share values. If share values (duly inflated by pumping money into the asset market) are a determinant of pay levels, then it is not surprising that the increasing ratio has manifested. This relationship is clear:
The answer to the so-called shareholder value problem – causing income inequality and low investment levels – is clearly to return to basics. The Fed and other central banks must reduce the assets on their balance sheets and stop pumping money into the stock market, so that those seeking investment destinations for their funds will invest it in productive plant with real prospects of wealth creation.
The problem of shareholder maximisation is sought to be addressed by the authors by making an appeal to the good natures of capitalist shareholders and boards of directors. That will not work. One might as well ask workers in a socialist state to work hard because it is good for the poor people in society. It has never worked, never will. People respond to incentives. If shareholder value maximisation is a problem in the sense that profits are not reinvested in businesses, then incentives to invest in productive processes must be restored. That would firstly require that the incentives for the current alternative, namely to invest in paper assets, must be removed. Real productive investment in the form of factory plant, information technology and worker training must become a competitive alternative again by eliminating the easy, rent-seeking alternative of share pyramids on the back of asset inflation.
In conclusion then, I do not agree with the diagnosis of the disease of stagnation in the so-called ‘neoliberal’ economies. In addition to targeting consumer inflation, there should also be targeting of asset inflation. Training and education should be beefed up. Labour markets and international trade should be freed up further. A good model to follow here is Switzerland. From a policy point of view, Switzerland is number four on the list of the freest economies in the world. It has good education, relatively low government spending for a rich country, low levels of labour regulation and union density and low tariff rates. It has high levels of basic, VET and tertiary training, high levels of investment, a strong non-inflationary currency, restricted state spending and a huge positive trade balance, and moderate and declining inequality of income.
Where the developed world has gone wrong, is by growing state expenditure and regulation, and by giving the monetary printing presses free reign. It can all be fixed, but not by means of this protectionist, old-style unionist, neo-Keynesian recipe for disaster.