Does providing information about the costs and benefits of automation affect the perceived fairness of a firm's decision to automate or support for government policies addressing automation's labor market consequences? To answer these questions, we use data from vignette and conjoint experiments across four advanced economies (Australia, Canada, the UK, and the US). Our results show that despite people's relatively fixed policy preferences, their evaluation of the fairness of automation—and therefore potentially the issue's political salience—is sensitive to information about its trade-offs, especially information about price changes attributable to automated labor. This suggests that the political impact of automation may depend on how it is framed by the media and political actors.

We investigate the numbers of complex zeros of Littlewood polynomials $p\left( z \right)$ (polynomials with coefficients {−1, 1}) inside or on the unit circle $\left| z \right|\,=\,1$, denoted by $N\left( p \right)$ and $U\left( p \right)$, respectively. Two types of Littlewood polynomials are considered: Littlewood polynomials with one sign change in the sequence of coefficients and Littlewood polynomials with one negative coefficient. We obtain explicit formulas for $N\left( p \right)$, $U\left( p \right)$ for polynomials $p\left( z \right)$ of these types. We show that if $n\,+\,1$ is a prime number, then for each integer $k,\,0\,⩽\,k\,⩽\,n-1$, there exists a Littlewood polynomial $p\left( z \right)$ of degree $n$ with $N\left( p \right)\,=\,k$ and $U\left( p \right)\,=\,0$. Furthermore, we describe some cases where the ratios $N\left( p \right)/n$ and $U\left( p \right)/n$ have limits as $n\,\to \,\infty $ and find the corresponding limit values.

We show that if a Barker sequence of length $n>13$ exists, then either n $=$ 3 979 201 339 721 749 133 016 171 583 224 100, or $n > 4\cdot 10^{33}$. This improves the lower bound on the length of a long Barker sequence by a factor of nearly $2000$. We also obtain eighteen additional integers $n<10^{50}$ that cannot be ruled out as the length of a Barker sequence, and find more than 237 000 additional candidates $n<10^{100}$. These results are obtained by completing extensive searches for Wieferich prime pairs and using them, together with a number of arithmetic restrictions on $n$, to construct qualifying integers below a given bound. We also report on some updated computations regarding open cases of the circulant Hadamard matrix problem.

Let $\text{ }\!\!\lambda\!\!\text{ }\left( n \right)$ denote the Liouville function. Complementary to the prime number theorem, Chowla conjectured that

*

$$\sum\limits_{n\le x}{\lambda \,\left( f\left( n \right) \right)}=o\left( x \right)$$

for any polynomial $f\left( x \right)$ with integer coefficients which is not of form $bg{{\left( x \right)}^{2}}$.

The classical theorem of Müntz and Szász says that the span of

is dense in C[0,1] in the uniform norm if and only if . We prove that, if {λi} is lacunary, we can replace the underlying interval [0,1] by any set of positive measure. The key to the proof is the establishment of a bounded Remez-type inequality for lacunary Müntz systems. Namely if A ⊆ [0,1] and its Lebesgue measure µ(A) is at least ε > 0 then

where c depends only on ε and Λ (not on n and A) and where Λ:=infiλi+1/λi>1.

Let ${{A}_{n}}\,=\,\{{{a}_{0}}+{{a}_{1}}z+\,.\,.\,.\,+\,{{a}_{n-1}}{{z}^{n-1}}\,:\,{{a}_{j}}\,\in \,\{0,1\}\}$ , whose elements are called zero-one polynomials and correspond naturally to the ${{2}^{n}}$ subsets of $\left[ n \right]\,:=\,\{0,\,1,\,.\,.\,.\,,\,n-1\}$. We also let ${{A}_{n,m\,}}\,=\,\{\alpha \left( z \right)\,\in \,{{A}_{n}}\,:\,\alpha \left( 1 \right)\,=\,m\}$ , whose elements correspond to the $\left( _{m}^{n} \right)$ subsets of $\left[ n \right]$ of size $m$, and let ${{B}_{n}}\,=\,{{A}_{n+1}}\,\backslash \,{{A}_{n}}$ , whose elements are the zero-one polynomials of degree exactly $n$.

Many researchers have studied norms of polynomials with restricted coefficients. Using $\|\alpha {{\|}_{p}}$ to denote the usual ${{L}_{p}}$ norm of $\alpha$ on the unit circle, one easily sees that $\alpha \left( z \right)\,=\,{{a}_{0}}+{{a}_{1}}z+.\,.\,.+{{a}_{N}}{{z}^{N}}\,\in \,\mathbb{R}\left[ z \right]$ satisfies $\|\alpha \|_{2}^{2}\,=\,{{c}_{0}}$ and $\|\alpha \|_{4}^{4}\,=\,c_{0}^{2}\,+\,2\left( c_{1}^{2}\,+\,.\,.\,.\,+\,c_{N}^{2} \right)$ , where ${{c}_{k}}\,:=\,\sum _{j=0}^{N-k}\,{{a}_{j}}{{a}_{j+k}}\,\text{for}\,\text{0}\le \,\text{k}\,\le N$.

If $\alpha \left( z \right)\,\in \,{{A}_{n,m}}$ , say $\alpha \left( z \right)\,=\,{{z}^{{{\text{ }\!\!\beta\!\!\text{ }}_{1}}}}\,+\,.\,.\,.\,+\,{{z}^{{{\text{ }\!\!\beta\!\!\text{ }}_{m}}}}$ where ${{\beta }_{1}}\,<\,.\,.\,.\,<\,{{\beta }_{m}}$ , then ${{c}_{k}}$ is the number of times $k$ appears as a difference ${{\beta }_{i}}\,-\,{{\beta }_{j}}$ . The condition that $\alpha \,\in \,{{A}_{n,m}}$ satisfies ${{c}_{k}}\,\in \,\{0,1\}$ for $1\,\le \,k\,\le \,n\,-\,1$ is thus equivalent to the condition that $\{{{\beta }_{1}},\,.\,.\,.\,,\,{{\beta }_{m}}\}$ is a Sidon set (meaning all differences of pairs of elements are distinct).

In this paper, we find the average of $\left\| \alpha \right\|_{4}^{4}$ over $\alpha \,\in \,{{A}_{n}}$, $\alpha \,\in \,{{B}_{n}}$ and $\alpha \,\in \,{{A}_{n,m}}$. We further show that our expression for the average of $\left\| \alpha \right\|_{4}^{4}$ over ${{A}_{n,m}}$ yields a new proof of the known result: if $m\,=\,o\left( {{n}^{1/4}} \right)$ and $B\left( n,\,m \right)$ denotes the number of Sidon sets of size $m$ in $\left[ n \right]$, then almost all subsets of $\left[ n \right]$ of size $m$ are Sidon, in the sense that ${{\lim }_{n\to \infty }}\,B\left( n,\,m \right)/\left( _{m}^{n} \right)\,=\,1$.

The main result of this paper gives an explicit computation of the $L_4$ norm of any cyclotomic polynomial of the form \[ f(x)=\Phi_{p_1}(\pm x)\Phi_{p_2}(\pm x^{p_1})\cdots \Phi_{p_r}(\pm x^{p_1 p_2\cdots p_{r-1}}). \] Here $\Phi_{p}$ is the $p$th cyclotomic polynomial and the $p_i$ are primes that are not necessarily distinct.

A corollary of this is the following theorem.

THEOREM. If\[ f(x)=\Phi_{p_1}(\e_1 x)\Phi_{p_2}(\e_2 x^{p_1})\cdots \Phi_{p_r}(\e_r x^{p_1 p_2\cdots p_{r-1}}), \]where N = p1p2···prand εj = ±, then \[ \frac{\| f\|_4^4}{N^2} & \ge & \frac{\big\|\Phi_2(-x)\Phi_2(-x^2)\cdots \Phi_2\big({-}x^{2^{r-1}}\big) \big\|_4^4}{4^r}\\& = & \frac{\big(\frac12 +\frac{5}{34}\sqrt{17}\big)(1+\sqrt{17})^r - \big({-}\frac12 +\frac{5}{34}\sqrt{17}\big)(1-\sqrt{17})^r}{4^r}. \] In particular the minimum possible normalized L4 norm of any polynomial of the above form is attained by \[ \Phi_2(-x)\Phi_2(-x^2)\cdots \Phi_2\big({-}x^{2^{r-1}}\big). \]

The minimum value of the Mahler measure of a nonreciprocal polynomial whose coefficients are all odd integers is proved here to be the golden ratio. The smallest measures of reciprocal polynomials with $\pm1$ coefficients and degree at most 72 are also determined.

This paper studies the spectrum that results when all height one polynomials are evaluated at a Pisot number. This continues the research theme initiated by Erdős, Joó and Komornik in 1990. Of particular interest is the minimal non-zero value of this spectrum. Formally, this value is denoted as $l^1(q)$ , and this definition is extended to all height $m$ polynomials as \[ l^m(q):= \inf(\vert y\vert: y = \epsilon_0 + \epsilon_1 q^1 + \cdots + \epsilon_n q^n,\, \epsilon_i \in {\bb Z},\, \vert\epsilon_i\vert \leqslant m,\, y \ne 0). \] A recent result in 2000, of Komornik, Loreti and Pedicini gives a complete description of $l^m(q)$ when $q$ is the Golden ratio. This paper extends this result to include all unit quadratic Pisot numbers. A main theorem is as follows.

THEOREM. Let $q$ be a quadratic Pisot number that satisfies a polynomial of the form $p(x) = x^2 - ax \pm 1$ , with conjugate $r$ . Let $q$ have convergents $\{C_k/D_k\}$ and let $k$ be the maximal integer such that \[ \vert D_k r - C_k\vert \leqslant m \frac{1}{1-\vert r\vert}; \] then \[ l^m (q) = \vert D_k q - C_k \vert. \]

A value related to $l(q)$ is $a(q)$ , the minimal non-zero value when all ${\pm}1$ polynomials are evaluated at $q$ . Formally, this is \[ a(q) := \inf(\vert y\vert: y = \epsilon_0 + \epsilon_1 q^2 + \cdots + \epsilon_n q^n,\, \epsilon_i = {\pm} 1,\, y \ne 0). \] An open question concerning how often $a(q) = l(q)$ is also answered in this paper.

We give explicit formulas for the ${{L}_{4}}$ norm (or equivalently for the merit factors) of various sequences of polynomials related to the polynomials

$$f\left( z \right):=\,\sum\limits_{n=0}^{N-1}{\left( \frac{n}{N} \right){{z}^{n}}.}$$

and

$${{f}_{t}}(z)\,=\,\sum\limits_{n=0}^{N-1}{\left( \frac{n+t}{N} \right){{z}^{n}}.}$$

where $\left( \frac{.}{N} \right)$ is the Jacobi symbol.

Two cases of particular interest are when $N\,=\,pq$ is a product of two primes and $p\,=\,q\,+\,2$ or $p\,=\,q\,+\,4$. This extends work of Høholdt, Jensen and Jensen and of the authors.

This study arises from a number of conjectures of Erdős, Littlewood and others that concern the norms of polynomials with −1, 1 coefficients on the disc. The current best examples are of the above form when $N$ is prime and it is natural to see what happens for composite $N$.

Let q be a prime and χ be a non-principal character modulo q. Let

formula here

where 1 [les ] t [les ] q is the character polynomial associated to χ (cyclically permuted t places). The principal result is that for any non-principal and non-real character χ modulo q and 1 [les ] t [les ] q,

formula here

where the implicit constant is independent of t and q. Here ∥·∥4 denotes the L4 norm on the unit circle.

It follows from this that all cyclically permuted character polynomials associated with non-principal and non-real characters have merit factors that approach 3. This complements and completes results of Golay, Høholdt and Jensen, and Turyn (and others). These results show that the merit factors of cyclically permuted character polynomials associated with non-principal real characters vary asymptotically between 3/2 and 6.

The averages of the L4 norms are also computed. Let q be a prime number. Then

formula here

where the summation is over all characters modulo q.

We consider the problem of minimizing the uniform norm on $[0, 1]$ over non-zero polynomials $p$ of the form

$$p(x) = \sum_{j=0}^n a_jx^j \quad\text{with } |a_j| \le 1,\, a_j \in {\Bbb C},$$

where the modulus of the first non-zero coefficient is at least $\delta > 0$. Essentially sharp bounds are given for this problem. An interesting related result states that there are absolute constants $c_1 > 0$ and $c_2 > 0$ such that

$$ \exp (-c_1 \sqrt{n} ) \le \inf_{0 \ne p \in {\cal F}_n }\|p\|_{[0, 1]} \le \exp (-c_2 \sqrt{n}),$$

for every $n \ge 2$, where ${\cal F}_n$ denotes the set of polynomials of degree at most $n$ with coefficients from $\{-1, 0, 1 \}$. This Chebyshev-type problem is closely related to the question of how many zeros a polynomial from the above classes can have at $1$. We also give essentially sharp bounds for this problem.

{\em Inter alia} we prove that there is an absolute constant $c > 0$ such that every polynomial $p$ of the form

$$p(x) = \sum_{j=0}^n a_j x^j, \quad\text{with } |a_j| \le 1,\,|a_0|=|a_n|=1,\, a_j \in {\Bbb C},$$

has at most $c\sqrt{n}$ real zeros. This improves the old bound $c\sqrt{n\log n}$ given by Schur in 1933, as well as more recent related bounds of Bombieri and Vaaler, and, up to the constant $c$, this is the best possible result.

All the analysis rests critically on the key estimate stating that there are absolute constants $c_1 > 0$ and $c_2 > 0$ such that

$$|f(0)|^{c_1/a} \le \exp (c_2/a)\|f\|_{[1-a,1]},$$

for every $f \in {\cal S}$ and $a \in (0, 1]$, where $\cal S$ denotes the collection of all analytic functions $f$ on the open unit disk $D := \{ z \in {\Bbb C}: |z| < 1 \}$ that satisfy

$$|f(z)| \le \frac{1}{1-|z|}\quad \text{for } z \in D.$$

We are interested in how small a root of multiplicity k can be for a power series of the form f(z:= 1+[sum ]∞n=1aizi with coefficients ai in [−1, 1]. Let r(k) denote the size of the smallest root of multiplicity k possible for such a power series. We show that

formula here

We describe the form that the extremal power series must take and develop an algorithm that lets us compute the optimal root (which proves to be an algebraic number). The computations, for k[les ]27, suggest that the upper bound is close to optimal and that r(k)∼1−c/(k+1), where c=1.230….

For a fixed algebraic number α we discuss how closely α can be approximated by a root of a {0, +1, -1} polynomial of given degree. We show that the worst rate of approximation tends to occur for roots of unity, particularly those of small degree. For roots of unity these bounds depend on the order of vanishing, k, of the polynomial at α.

In particular we obtain the following. Let BN denote the set of roots of all {0, +1, -1} polynomials of degree at most N and BN(α k) the roots of those polynomials that have a root of order at most k at α. For a Pisot number α in (1, 2] we show that

and for a root of unity α that

We study in detail the case of α = 1, where, by far, the best approximations are real. We give fairly precise bounds on the closest real root to 1. When k = 0 or 1 we can describe the extremal polynomials explicitly.

Sharp extensions of some classical polynomial inequalities of Bernstein are established for rational function spaces on the unit circle, on K = r (mod 2 π), on [-1, 1 ] and on ℝ. The key result is the establishment of the inequality

for every rational function f=pn/qn, where pn is a polynomial of degree at most n with complex coefficients and

with | aj | ≠ 1 for each j and for every zo∈ δ D, where δ D,= {z∈ ℂ: |z| = l}. The above inequality is sharp at every z0∈δD.

For a sequence of polynomials (Pn) orthonormal on the interval [−1, 1], we consider the sequence of transforms (gn) of the series given by . We establish necessary and sufficient conditions on the matrix (bnk) for the sequence (gn) to converge uniformly on compact subsets of the interior of an appropriate ellipse to a function holomorphic on that interior.

The operator Im is defined as m-fold indefinite integration with zero constants of integration. The zero distribution of Im(p) for polynomials p is studied in general, and for two special classes of polynomials in detail. The main results are: (i) The zeros of In(Pn), where Pn(𝑧) is the n-th Legendre polynomial, converge to a certain algebraic curve; (ii) the zeros of an integer) converge to pieces of a circle and of two "Szegö curves".